Safari in Mavericks: PDF downloading when native display is off is broken

Posted by Pierre Igot in: Macintosh
October 23rd, 2013 • 5:56 pm

I have been posting small notes about my experience with OS X 10.9 (Mavericks) on my Twitter feed, but this one warrants a longer post — if only to properly express my indignation.

For months now, I have been testing the newest version of Safari 6 as part of my AppleSeed membership. And right away, I noticed one particular issue that affects people who, like me, do not like the so-called “native” display of PDF files inside the browser window and would rather get all their PDF files automatically downloaded to their “Downloads” folder instead. (See this older post for more on this.)

The newest version of Safari does not eliminate the ability to disable native PDF display via a command in Terminal. However, it breaks it in a way that makes it useless.

Here’s what now happens in Safari after you disable native PDF display and click on a PDF to download it:

Safari7-PDFdisplaybroken

It displays the PDF as raw data inside the browser window!

And of course, Safari being what it is, if the PDF in question happens to be fairly large, the app has a strong tendency to become totally unresponsive at the same time, with the typical “Web pages are not responding” alert, which obliges you to reload all your currently open pages in all your Safari tabs and windows. (How this is better than the Safari app simply crashing, I do not know. Apparently, it’s an improvement.)

What really infuriates me about this particular bug is that I noticed it as soon as I installed the very first seed of the new Safari, several months ago. And I immediately filed a bug report. I tried to be patient, but when I saw that the bug was not fixed in subsequent seeds, I tried submitting more reports.

Not only did they never acknowledge my initial report, but they also never flagged my subsequent reports as “duplicate”. It was like talking to a wall. At that stage, I became reasonably confident that they would fail to fix the bug in the final release, and sure enough, they did not fix it and now everyone can enjoy it as much as I have been doing for the past few months.

Of course, you can avoid the buggy behaviour by option-clicking on links to PDF files in order to force Safari to download the files instead of displaying them (wait until they break that feature in the next version of Safari!), but:

  1. It is a pain in the neck to have to remember to hold the option key down when clicking on a link to a PDF file, especially when it is not obvious that it is a link to a PDF file and not to another web page.
  2. This option-clicking workaround works even when native PDF display is enabled, so the ability to disable native PDF display is now effectively useless. It might as well not exist at all.

Apple cares very little about the needs of so-called “power users” these days, so I am not entirely surprised that they failed to acknowledge my bug reports and fix the bug in time for the final release, and I am not too optimistic that they will fix it any time soon — although the fact that the same feature was already broken once in the past and eventually fixed by Apple back then (see this older post) gives me a bit of hope (although it took many months). But it would help if everyone affected could bombard Apple with bug reports now that the new Safari is out.


Word 2011: Issue with text selection in right-aligned table cells

Posted by Pierre Igot in: Microsoft
October 23rd, 2013 • 10:21 am

In Word 2011, if you have a table with cells where the text alignment setting is “Align Text Left”, and if your cell in empty, when the cell is selected, the selection highlighting and the mouse pointer look like this:

word2011-tablecell-leftaligned-selected

If you then click once anywhere in the cell, Word 2011 deselects the empty cell and put the I-beam cursor instead:

word2011-tablecell-leftaligned-clicked

On the other hand, if the text alignment setting for the cell is “Align Text Right”, and if your cell in empty, when the cell is selected, the selection highlighting and the mouse pointer look like this:

word2011-tablecell-rightaligned-selected

If you then try to click once anywhere in the cell, Word 2011 refuses to deselect the empty cell and put the I-beam cursor instead. In order to achieve the same effect, your only option is to switch to the keyboard and press a cursor key. Then finally Word 2011 deselects the empty cell and put the I-beam cursor instead:

word2011-tablecell-rightaligned-clicked

This problem only occurs when the cell is empty. If it already contains some right-aligned text, you can click on the text to deselect it and get the I-beam cursor instead.

There are some many little “details” like this one that are not right in Word 2011. And no one seems to care. It’s sad.


So I just had brain surgery…

Posted by Pierre Igot in: Society, Technology
September 15th, 2013 • 2:09 pm

I don’t usually share personal stuff here, but it went well, and it’s done, so I feel fortunate to be able to talk about it and share some of my (slightly amputated) thoughts on the whole process, especially when it comes to the “user-friendliness” of the thing, and boy, they sure could use some help in that department.

The context in a nutshell: Through a routine IRM of my head a few years ago, a suspicious area was spotted in my brain, above the right eye. I didn’t have any obvious symptoms of anything, so the decision was made to simply monitor the thing with regular IRMs and revisit the decision if necessary.

Three years later, there appeared to be enough of a change to justify intervention and so… under the knife I went. The tumour was removed on Thursday and I am here at home typing at my computer, so obviously things went quite well as far as the surgery itself is concerned (although of course I am a bit, let’s say, stunned and will still need to pace myself for quite a while).

The neurosurgeron is reasonably confident that he got “the whole thing” out (as far as that is possible), as was confirmed by the IRM that I had the day after the surgery, but we have to wait for the results of the biopsy to see what kind of tumour we are actually dealing with (there are many kinds, apparently). However, chances are that it is of the “low grade, slow-growing” kind, which will require on-going monitoring for the rest of my life, and hopefully not much else.

As for the process, a sure sign that my brain was functioning normally after the operation was that I immediately started thinking about all the things that could probably be easily improved in the recovery process, especially when it comes to technology (of course).

One of the most annoying aspects of being stuck in a “step-down” ward with other recent victims of surgery is that you are all hooked up to all kinds of monitoring devices that emit all kinds of signals. Easily the most annoying one was the arterial hook-up monitoring the heart rate. The thing is obviously configured by default to beep every time there is a sudden change in your heart rate. But apparently this happens so often that nurses have learned to ignore the beeps, unless of course they become persistent. (They can change the settings, but I suspect it’s too much of an annoyance to do it systematically and so… You know the rest.)

The stupidest aspect, as far as I am concerned, is that, at least in my case, the simple fact of falling asleep apparently causes enough of a change in my heart rate to trigger the alarm. As you can imagine, falling asleep is not the easiest thing to do in a busy ward at the best of time. When you’ve been put to sleep for six hours and have a bit of a headache, it’s even harder. But really, it does not help at all that, right at the time when you finally fall asleep, the idiotic heart rate monitor emits a beep that is, of course, close enough to wake you from your not-quite-asleep-yet state.

Next on the list are the IV pumps. These things sound exactly like inkjet printers going about their business or, worst-still, dot-matrix printers from the 1980s. Given that they have to be running permanently, you’d think they would put more effort into coming up with quieter designs. Yes, it is a repetitive sound and you eventually get used to it, but surely it’s one more thing that it should not be necessary to inflict on patients trying to regain their strength and get some much needed sleep.

Then of course you add the usual rudeness and general impoliteness of other people, including other patients and their visitors. For the first night after the surgery, my four-bed ward was, from my perspective, a Bosch painting come to life. Between the alcoholic/addict woman that had been hit by a car, with a broken hip and pelvis and something to her head, who kept pulling her tubes out and swearing in some incomprehensible language (there were enough “shits” and “fucks” in there for me to be able to tell that she was swearing), the aneurysm patient who had been driving all the nurses crazy for ten days straight by refusing to comply with their orders and attempting to get up and go away, himself releasing an endless stream of “oh my fucking god”s and other choice words, not to mention innumerable burps and farts and what not, and then the poor old soul who showed up at 5 am with a broken neck that, with his history of lung cancer, prevented him from breathing properly, instead making sounds that clearly indicated that he was drowning in his own phlegm and requiring the nurses to suction stuff out on a regular basis, all with nice, appealing noises for the rest of the room as you can imagine, it was quite a scene. Of course, I am not ignoring the fact that the people moaning and swearing were in real pain and clearly the morphine drips were enough to shut them up for a while, but it was never a very long while and I just did my best to enjoy my own morphine silently and keep my farts low-key.

The surgeon got word of the general ambience of the place and was kind enough to have me moved to a different ward for the next night, which was less apocalyptic, but the IV pumps were still active in their sonic ImageWriter-like majesty. (Fortunately, no one was hooked up to a heart rate monitor any longer at that stage.)

And then I was allowed to go home, which is of course a much better place to recover, under the care of my wife and with my familiar luxuries. I am not supposed to stay at the computer for too long, so I think I’ll keep it at that. I just sort of wanted to explain why things had been rather quiet for me on the blog/Twitter front lately.


Adobe InDesign CC: Selection colour and responsiveness affected by transparency of linked PSD file

Posted by Pierre Igot in: Macintosh
August 14th, 2013 • 9:05 am

Yesterday, I wrote a post about a new problem introduced in InDesign CC that involves the presence of a link to a PSD file on the page.

Thanks to the feedback of Betalogue reader and fellow InDesign user Lars P., I am now able to confirm that the problem is more specifically linked to the use of transparency. As Lars says:

As soon as there are transparent objects (placed images, drop shadows, …) on a page, InDesign renders the whole page using the transparency blend space. That is normally set to CMYK, so all colours get converted to your CMYK working space ICC profile for display. Most vibrant colours (like your green for highlighting) will look washed out then.

Sure enough, changing the transparency blend space to “Document RGB” as opposed to “Document CMYK” does change the selection highlighting colour back to its normal shade. It does not, however, address the responsiveness issue that is also associated with this particular problem. (And you also need to remember to change it back to CMYK before sending the publication to the printer, if that’s part of your process.)

As I indicated yesterday, removing the link to the Photoshop file from the page eliminates the problem altogether. But that’s not necessarily an option. I have found that, if you do not need transparency and can flatten the Photoshop document, this also causes InDesign CC to stop rendering the whole page using the transparency blend space. The selection highlighting colour changes back to its normal shade, and the responsiveness issue is solved.

In other words, both the change in the selection highlighting colour and the responsiveness issue appear to be due to the use of the transparency blend space by InDesign CC to render the page. According to Lars, that behaviour with transparency blend space and changing colours has been present in InDesign since version 2. But I certainly have never noticed it before InDesign CC on my machine with the kind of publications that I work on, so I suspect that Adobe did change something in InDesign CC that now causes the problem to surface in situations where it did not before.

I wouldn’t really care about the colour changes if the problem did not also affect the responsiveness of the application to mouse movements. But the responsiveness issue is very noticeable as far as I am concerned, and I will have to do everything that I can to avoid it. If it involves eliminating the transparency in the linked PSD file, that’s what I’ll do.

In this particular case, the transparency was not needed, as the background colour of the document was white, which was also the background colour of the PSD document involved once it was flattened. But I can easily imagine that I will soon encounter situations where I have no option but keep links to transparent PSD documents. And so that might force me to have to deal with the responsiveness issue.

Based on what Lars is saying, and the subjectivity of something such as responsiveness (it really depends on how fast you work with your mouse), I have little hope that Adobe’s engineers will notice the problem and do something about it. So it’s yet another flaw that one is going to be forced to live with… Sigh.

Thanks to Lars for his feedback!


Adobe InDesign CC: Presence of PSD graphic affects selection colour and responsiveness

Posted by Pierre Igot in: Macintosh
August 13th, 2013 • 3:38 pm

It’s an inevitability: Every new version of a major software title introduces new bugs. I’ve just encountered one in InDesign CC, and it didn’t take an inordinate amount of sleuthing to identify its source.

The problem surfaced when I started working on a publication designed by a client, which I am expected to translate, with the required adjustments to the layout for the French text. I have been doing this for this particular client for a couple of years now on various publications in the same series, so the files in question were nothing new.

What was new in InDesign CC, however, was that sometimes the application switched from my regular selection highlighting colour (a dark green) to a different shade.

Here’s what my normal selection colour looks like:

InDesignCC-PSDfile0

And here’s what it looked like in some situations:

InDesignCC-PSDfile1

The problem was not purely cosmetic, however. The selection colour was different, but there was also a noticeable degradation in selection highlighting performance. In other words, when I dragged the mouse around to extend or shorten the current selection, there would be a small delay between my movements and the corresponding changes on the screen.

I am someone who works very fast, and this delay was very noticeable to me, making me consistently miss my targets because the screen display was no longer in sync with my mouse movements. It was very irritating.

I started exploring InDesign CC’s settings, to see if there was something about the display performance options that had changed. But I couldn’t find anything obvious. I also quickly noticed that the problem seemed to come and go as I was working on my publication, for no visible reason.

I also noticed that, if I scrolled so that the section of the document with the selection highlighting was off-screen, and then scrolled back to make it visible again, until I released the mouse button the selection colour was as expected:

InDesignCC-PSDfile2

As you can see, when you are scrolling in an InDesign window, the scrolling might be “live”, but the application only draws the content revealed by the scrolling partially, by using a lower-quality version of the text and graphics, without anti-aliasing. The anti-aliasing is only added when you release the mouse button to indicate that the scrolling is done. Here’s a screen shot that better illustrates the trick:

InDesignCC-PSDfile5

(In this screen shot, I am bringing back the right half of the text back into view by scrolling to the right.)

This is presumably done in order to offer better scrolling performance, also it seems like a rather antiquated mechanism to me. (And text and graphics without anti-aliasing sure look ugly — although Adobe’s own anti-aliasing is not the best looking one in the neighbourhood either!)

So this led me to suspect that the problem might have something to do with anti-aliasing. But turning anti-aliasing off altogether didn’t fix the problem — somewhat to my relief, since having to work with anti-aliasing would definitely have hurt my eyes.

I was back to square one. Then, while I continued working on my publication, I noticed that the problem only seemed to affect certain pages in the publication. So it was not randomly intermittent. It was intermittent because there was something on some pages that caused the problem to occur.

I ended up narrowing it down by creating a copy of the publication and deleting a bunch of stuff. Within a few minutes, I was able to establish that the problem was most definitely caused by the presence of a link to a Photoshop document (a “.psd” file) on the page. As soon as I deleted the link, the selection highlighting colour reverted back to the expected shade, and the normal selection highlighting responsiveness was back. As soon as the link to the Photoshop file was restored, the problem reappeared.

I was able to reproduce this in a blank document with a link to another Photoshop file, so it does not seem to be linked to a particular Photoshop file or to other aspects of the publication.

So here we are… A new bug to add to the list. For what it’s worth, I have reported it via this page.


Fix for kernel panics on Mac Pro with multiple GeForce cards: ATI Radeon HD 5770

Posted by Pierre Igot in: Macintosh
August 1st, 2013 • 9:08 am

In the summer of 2012, Apple started seeding early builds of the OS X 10.8.2 update. As a member of the AppleSeed program, I was able to test these early builds. I quickly noticed (as one would) that they introduced a serious new bug that caused frequent kernel panics on my Mac Pro, up to several times a day. I of course immediately filed a bug report on these kernel panics.

In September 2012, the final version of the OS X 10.8.2 update came out, and the kernel panics were still there. Around the same time, Apple sent me a response to my bug report indicating that the problem was a “known issue” and closing my bug report as a “duplicate”. Meanwhile, a thread discussing the issue appeared on the Apple forums, and it quickly became apparent that the problem could be narrowed down to people running Mac Pro computers with multiple GeForce video cards driving multiple monitors, and involved a few kernel extension files that were updated in OS X 10.8.2.

In October 2012, Apple released an update called OS X 10.8.2 Supplemental Update, which updated one of the kernel extensions involved. Sadly, it quickly became apparent that the supplemental update did not fix the kernel panics.

In late 2012, Apple started seeding early builds of the OS X 10.8.3 update. It included updated versions of all three kernel extensions involved in the kernel panics. The update did appear to have an impact on the frequency of the kernel panics, at least on my machine. But it failed to eliminate them completely.

In March 2013, the final version of OS X 10.8.3 came out. Again, while it included updated versions of all three kernel extensions involved in the kernel panics, it failed to completely eliminate the kernel panics.

Meanwhile, Mac Pro users managed to identify various ways of dealing with the kernel panics. I myself provided a hack that offered temporary relief, at least for some people. But it was not a fix.

I purchased an expensive Mini DisplayPort to Dual-Link DVI Adapter, which enabled me to connect both of my Apple Cinema 30-inch displays to the same GeForce video card and stop using the second card. This eliminated the kernel panics altogether, but it also had a noticeable impact on video performance and general system responsiveness, undoubtedly due to the fact that my system now only had half as much VRAM and video card processing power available.

Some people in the Apple Discussions thread also mentioned that replacing one of the GeForce video cards with an ATI Radeon HD 5770 appeared to provide a permanent fix for the problem.

Personally, I went back to using my two GeForce cards when Apple started seeding OS X 10.8.3, which enabled me to continue to monitor the situation, but of course also meant that I continued to experience the kernel panics, albeit less frequently. This remained true with the final OS X 10.8.3 update, and with early builds of the OS X 10.8.4 update.

In June 2013, Apple released the final version of OS X 10.8.4, and it still did not fix the kernel panics. As I indicated at the time, it looked like Apple was working on the problem, especially since “graphics drivers” was listed as a focus area during the testing of early builds of the update. But the fact remained that the kernel panics were still there in the final version of OS X 10.8.4.

In spite of my willingness to voluntarily endure some level of computing hardship in the name of helping Apple improve its products, I finally decided that I had had enough, and purchased the ATI Radeon HD 5770 upgrade kit myself.

I installed it in early July, and I can now confirm, one month later, that it is indeed a permanent workaround for the problem. (I am reluctant to call it a “fix” myself, because the problem might be still there on the software side of things. But the upgraded hardware prevents OS X from encountering the software problem in the first place, so it’s as good as a fix.) I have not had a single kernel panic since installing the video card.

I can also confirm that installing the Radeon card is fairly straightforward and, for those (like me) sensitive to noise issues, I can also confirm that, even though the card is bulkier and more powerful and requires its own power supply, it has no substantial impact on the overall sound level produced by the Mac Pro. (Sound is a very subjective thing, and there are multiple factors involved. In the summer, for example, when it gets really hot, there is definitely more fan noise coming from my Mac Pro. But I don’t believe the situation is any different from what it was with the two GeForce cards.) Also, while the Radeon 5770 upgrade is listed for “Mac Pro (Mid 2010)”, it works just fine in my 2009 Mac Pro.

Of course, the card has double the VRAM of the GeForce and more processing power, so there is also a little bit of a boost performance-wise. But it’s the kind of small boost you very quickly get used to, and I don’t think I would say that it provides a significant improvement over driving the two Apple Cinema displays with two GeForce cards. It certainly does not appear to be worth purchasing if you are not having kernel panics to begin with.

The bottom line here is that Apple has failed Mac Pro users. While I cannot verify that the bug is still there in 10.8.5 (Apple has been releasing early builds for a while now), I wouldn’t be surprised if it turned out that it still is. How Apple can justify “knowing” about such a serious issue and not doing anything about it for several months or, indeed, nearly a year now, I just do not know.

As I said before, the cynical in me cannot help but wonder whether Apple is not just taking advantage of the fact that Mac Pro users (especially those with multiple monitors) tend to have more disposable income and don’t mind spending several hundred dollars to work around a problem created by the company in the first place. I know that, personally, being able to do my computing in an environment free of kernel panics is definitely worth the nearly $500 that I have now spent on this problem (not to mention the countless hours spent testing, bug reporting, and simply rebooting and restoring my work environment after crashes).

But it’s definitely not right, and Mac Pro users have every reason to be pissed off with Apple right now. (I am past that stage, but that’s mostly because I have taken steps to work around the problem and I have other important issues to worry about in my life these days.) Apple has let us down big time, and while there are still reasons to be excited about the new Mac Pro announced earlier this year, it will take more than a little effort on Apple’s part to repair its relationship with those affected by this very serious and (as yet?) unaddressed issue.


Delicious Library 3: A major disappointment

Posted by Pierre Igot in: Macintosh
July 18th, 2013 • 2:58 pm

I used to have a custom-made FileMaker Pro solution for cataloguing my books, my music, and my movies. But when Delicious Library came about, with its built-in access to Amazon’s data, I quickly made the decision to switch to this particular tool. The user interface had a few quirks, but on the whole the experience of using it for cataloguing and browsing my ever-expanding collection was a pleasant one.

There was even a golden era when Delicious Monster was actually able to provide an iOS application that would seamlessly sync with the desktop application and enable you to carry your entire collection around in a convenient portable form. Unfortunately, Amazon then imposed new restrictions that made it impossible for Delicious Monster to continue to offer the iOS application — as if it was “competing” with Amazon’s own on-line offerings as opposed to actually encouraging people to buy more stuff from Amazon. (The iOS application, while no longer available from the App Store, still works on my aging iPad and iPod touch devices, but it’s probably only a matter of time before I have to give up on it altogether.)

All this brings us to the major upgrade that is Delicious Library 3. I have no time to review all its new features, but I have to report that, for my own purposes and in my own daily use, the new version is a major disappointment.

One of the first things you will notice is that the application no longer has a “Preferences” dialog box. That tells you something about the flexibility and customizability that they have sacrificed, probably in the name of “simplifying” the user experience. There are still various viewing options, but these options are too limited and make the experience of using Delicious Library downright annoying.

For instance, the “View as List” option, which is the only one that enables you to view your collection in a workable format, now uses two lines for each item, which of course significantly reduces the number of items that can be seen at any given time. Worse still, you apparently can no longer sort this list by title or by creator alone. The only column option available is an option called “Title and Creator”:

Delicious Library-TitleandCreator

By default, this column uses a sorting rule that follows a rather non-obvious logic, separating those items that do have a creator from those who don’t have one. When you know how inconsistent Amazon’s info is, especially for movies, this means that, for example, various seasons of a single TV show on DVD (say, Breaking Bad) will end up being listed separately, because some seasons have something in their “Creator” field whereas other seasons do not. (And of course all these items that have nothing in their “Creator” field take up as much vertical space as those who do have something, so there’s lots of wasted screen real estate here.)

It seems to me that, since Delicious Library is so intimately linked to Amazon’s data sets, using a data structure that assumes that all the data sets are in perfect condition, with all fields properly filled out, is rather ridiculous.

In addition, my initial impression was that the “Creator” field always took priority over the “Title” field, which meant that it was effectively impossible to have a list of all your titles in pure alphabetical order. DL would always sort by creator first, and then by title. It was simply unacceptable.

I only recently discovered, quite by accident, that you can actually somehow force Delicious Library to sort by title first and then by creator, by clicking on the “Title” word itself in the column header — although I am still not sure exactly what combination of clicks is required. (Now that I have achieved this, I am afraid to undo it.) Somehow, after clicking multiple times, I managed to have a column header that only has a triangle (indicative of sort order) next to “Title” and not to “Creator”:

DeliciousLibrary-TitlesansCreator

Now finally my lists are only sorted by title, whether or not the items have something in their “Creator” field. Don’t ask me exactly how I did it, though. It’s still very mysterious to me. It certainly does not meet the most basic tests of “discoverability”.

The sort order is still not perfect, however. Even though DL obviously has an algorithm for ignoring the definite article “The” in titles (as well as its equivalent in other languages, such as French), I still have a number of DVDs whose title starts with “The” that are listed under “T”, even though the next word does not start with a “T”. I have no idea why.

When DL 3 was first released, there was — believe it or not — absolutely no way to keep item details visible at all times. DL 3 did away with the “Details” pane altogether, and the details were only visible in one of those horrible “pop-up” windows, like in iCal/Calendar. Thankfully, there was enough of an outcry to force the developer to bring back something similar to the “Details” pane from DL 2, in the form of a separate “Details” window.

Unfortunately, in my experience, the implementation of this window is buggy. I quite often find myself in another OS X application with the “Details” window from DL 3 still visible in the foreground above the windows of my other application. I can fix the problem by switching to DL 3 and then back to my application, but really… This is a bit much.

The whole application has also become significantly more sluggish than DL 2 used to be. Even on my relatively fast 2009 Mac Pro with 12 GB of RAM and an SSD for the system volume, everything is slow and choppy, including such basic things as scrolling up and down the list. (Let’s not even mention what happens in “View as Shelf”.) My collection is not tiny, but it’s not enormous either. I find such performance levels barely tolerable.

The search feature is not any better. While it has been fine-tuned somewhat since the original DL 3 release, you are still afraid of making a typo while typing your search request, because this will cause DL 3 to initiate a useless “search-as-you-type” search, which locks up the UI for several seconds, before you’ve finished typing your search request. Ugh.

The process of adding new items to the library is also affected by this sluggish performance. While this process has been simplified and you can now just scan your item’s barcode without bringing up a dialog first, if you are unfortunate enough to attempt to scan the barcode while the application’s UI is locked up, the application emits a useless system beep instead of buffering your data entry and processing it when it becomes responsive again. This happens to me every time I add a new item, because, once DL 3 has recognized the barcode and downloaded the data from Amazon, it takes a couple of seconds to move the newly-added item to the appropriate position in the sorted list of titles. If I happen to scan the next barcode during that resorting, all I get is a beep. Each time I add a new item, I have to wait until DL 3 has moved it to the appropriate place before scanning the next barcode. It’s quite frustrating. I frankly would rather have a modal dialog box than this, because effectively it’s the whole UI that is now modal in an invisible way and forces you, without any visual feedback, to wait until it’s finished “processing” the new item.

This whole situation makes me sad. Delicious Library has not become entirely unusable, but it’s definitely taken a huge step back in usability. I waited for a few months, hoping that these issues would be addressed in subsequent DL 3 updates, but there have been many updates now, and most of the issues I describe here are still there.

If this were a proper review, I would of course offer a more balanced report, by describing the new features, some of which are quite nice (such as the “Recently Added” list on the side). But for me, none of the new features comes anywhere near outweighing the flaws and limitations introduced in the new version. The sluggishness and the lack of viewing and sorting options are fundamental flaws that will have to be properly addressed before I find any pleasure in using this application again.


Pages ’09: ‘The file format is invalid’

Posted by Pierre Igot in: Microsoft, Pages
July 4th, 2013 • 4:15 pm

Ever got this error when you tried to open a Word document in Pages ’09?

The file format is invalid

I just did, and I now know why.

See, Microsoft Word users in general and Windows users in particular are so confused about file extensions, and Microsoft’s user interface for dealing with file extensions is so confusing, that sometimes they do things that Microsoft and Apple did not anticipate, like… saving a DOCX file with the “.doc” file extension.

That’s exactly what happened in this case. When I got the error with this particular “.doc” file and noticed that it was opening in Word 2011 just fine, I figured I’d try to change the file extension to “.docx” and see what would happen.

Sure enough, as soon as I changed the file extension, the renamed file opened just fine in Pages ’09.

In this day and age, you’d think that:

  1. Microsoft Word would be smart enough to detect the error and give a warning to the user to the effect that the file he’s opening is a DOCX file masquerading as a DOC file, and offer to fix the problem;
  2. Apple’s Pages ’09 would be smart enough to recognize the correct file format in spite of the wrong file extension and offer to open the file just the same, again with some kind of warning about the file extension.

But no… Both companies obviously have bigger fish to fry, and so we get no warning whatsoever in Microsoft Word, and a useless error message with no suggestion in Pages ’09.

As for smartness in computer software, I guess it’ll take another century or two.


OS X: Sandboxing = more crashes

Posted by Pierre Igot in: Macintosh
June 19th, 2013 • 9:09 am

This is something that pisses me right off.

Apparently, according to Apple, sandboxing is good for us. After all, it “provides a last line of defense against the theft, corruption, or deletion of user data if an attacker successfully exploits security holes in your app or the frameworks it is linked against”. Yay.

As far as I can remember, with no particular effort on my part to protect myself against thieves, hackers, and so on, I have never been the victim of a single act of theft, corruption, or deletion of my user data on my OS X by an attacker. (What has happened to me on-line is another matter.)

Since the introduction of this “sandboxing” thing, on the other hand, I have most definitely been the victim of this kind of thing:

mail-sandboxcrash

I am no expert, but I have little doubt that the above is a report on a crash that was caused by a bug in Apple’s “sandboxing” system (full crash log here). And it’s not the first one I have experienced. What did I do to deserve it? Nothing. I was simply composing an e-mail and tried to attach a file to it using the standard Open File dialog box.

So let me get this straight: In the name of improving my security, which has never been under threat as far as I can tell, Apple has introduced a new “feature” whose net effect is, quite simply, that I experience more crashes than I used to, through no fault of my own.

Great.

Now, let’s go back to the statement quoted above: sandboxing ““provides a last line of defense against the theft, corruption, or deletion of user data if an attacker successfully exploits security holes in your app or the frameworks it is linked against”.

Is it just me, or is this “last line of defense” not another way of telling developers: “Don’t worry if you introduce security holes in your app. We’ve got the user covered.” In other words, won’t this approach have the undesirable effect of making developers sloppier in their work?

And if the sandboxing system causes crashes so easily, how reliable can it really be anyway?

We had the nanny state. Now we have the nanny OS.


Word 2011: Why does word selection still select the trailing space?

Posted by Pierre Igot in: Microsoft
June 13th, 2013 • 10:11 am

In Microsoft Word, there is a myriad of inconsistencies, behaviours that don’t make sense and are simply the way they are because they have always been that way and nobody at Microsoft bothers to use his or her brains to think about whether these things still make sense today.

Take, for example, what happens when you double-click on a word in Word 2011. Double-clicking on a word is what is called “word selection”. No matter where you double-click on the word, it selects the entire word. It’s much faster and much less error-prone than having to position your cursor at the beginning of the word, to click-and-hold and then to drag your cursor to the end of the word. (And you can also select multiple words this way, by double-clicking on a word and holding the mouse button down while you continue to drag over several consecutive words. This is called “word-by-word selection”, as opposed to the default “character-by-character selection”.)

In every other OS X application, when you double-click on a word in the middle of a sentence, the application only selects that word. In Word 2011 (and every version before it), on the other hand, the application also selects the trailing space that comes after the word (if there is one, i.e. if the word is not immediately followed by a punctuation sign).

Why does Word 2011 do this? As far as I can tell, it’s because Word first came into being a long time ago, at a time when so-called “smart editing” features didn’t exist. These days, in most applications, when you double-click on a word to select it and then press Delete to delete it, the application is smart enough to also delete the extra trailing space that comes after it, so that you are not left with a situation where the two remaining words around the word that you’ve just deleted are separated by two spaces instead of one. Ditto if you cut the word with command-X.

This “smart” behaviour was introduced many years ago. Word itself introduced this feature many years ago. (If you don’t like it, you can turn it off in Word’s “Preferences” dialog box, under “Edit”, in the “Use smart cut and paste” customization options: the option is called “Adjust sentence and word spacing automatically” and it’s on by default.)

The thing is, the introduction of this very feature made the selection of the trailing space after the word irrelevant. As far as I can tell, the only reason why the original designers of Microsoft Word decided to include the trailing space with the word in word selection was in order to make it easier for writers to edit their text without having to constantly delete extra spaces. But once Word adopted the “smart” handling of word spacing, this automatic selection of the trailing space became irrelevant.

Try it: if, instead of double-clicking on a word to select it (and its trailing space) in a Word document, you select the word in question “manually” by positioning your cursor at the beginning of the word, clicking-and-holding and then dragging your cursor to the end of the word, without including the trailing space, and then you press Delete to delete the selected word, Word still deletes the extra space as well, because of the “Adjust sentence and word spacing automatically” option.

So why does word selection in Word still select the trailing space as well even today? As far as I can tell, it’s because no one at Microsoft has ever thought of using his or her brains to make the decision that, now that Word has smart handling of word spacing, the selection of the trailing space is no longer necessary.

Not only is it no longer necessary, but it is inconsistent with all other applications, and it has all kinds of undesirable side-effects. If you use word selection to select a word or a phrase in a Word document and then copy it to paste it in another application, the extra trailing space at the end will be copied and pasted along with it. Unless the destination application is an application that also has smart handling of word spacing (like another word processor), this extra space will remain.

It is the case, for example, if you use word selection to copy something in a Word document and then paste it in your web browser. You’ll always get that extra space along with it, and in many cases you’ll be forced to delete it manually in the destination application because it’s highly undesirable there.

Even within Word, the extra trailing space is highly undesirable when it comes to character-level formatting, because when you use word selection to select a word or a phrase and then apply some character-level formatting to the selection, this character-level formatting gets also applied to the trailing space, and so if, later on, you put your cursor after the trailing space to type something else, often times the new text you type will also have that character-level formatting, even though there is no visual indication that the space before it has that character-level formatting (a space in bold or italics looks just like a regular space). The worst situation is what happens with underlined text. I cannot count the times I see documents that contain underlined words where the underlining also extends to the space after the word, which is horrible from a typographic point of view.

But does anyone at Microsoft care about all this? Apparently not. Today, in 2013, many many years after smart spacing was introduced in Word itself, word selection in Word still includes the trailing space. It is yet another indication that there is no actual thinking about real-world usage that takes place at Microsoft in the minds of the software engineers working on Microsoft Word.


About the new Mac Pro

Posted by Pierre Igot in: Macintosh
June 11th, 2013 • 8:40 am

I am ambivalent about the upcoming new Mac Pro unveiled by Apple yesterday at the WWDC. Of course, I am relieved that Apple has finally given us something tangible to look forward to.

I am also glad that the new model is a radical rethink and that heavy emphasis has apparently been put on keeping the machine as quiet as possible (although of course how quiet it really will be remains to be seen, or rather heard).

I am happy that it will have lots of raw power. While my current machine (a 2009 Mac Pro) is no slouch, I am always looking for ways to increase my productivity. I am not one of those apparently numerous (if you believe the pundits) Mac users who feel that their current machine is fast enough. If you have ever tried to type text in a textbox in a Microsoft Word 2011 document or to resize a document window in Adobe InDesign, you know that there is still plenty of room for improvement. Since Microsoft’s and Adobe’s engineers obviously feel that their own software is “good enough”, we have no choice but to rely on the raw power of the underlying hardware to improve real-life performance in those applications.

I also appreciate the focus on raw graphics power. While I am not a videographer, I have lots of screen real estate (two 30-inch displays) and I wouldn’t mind having even more. A large Retina display is still probably several years away, but I might be tempted by a larger screen at some point, even if Apple itself does not seem interested in producing such screens anymore.

(I suppose that, when the time comes to buy one of these new Mac Pros, I’ll also have to spend some extra money on adapters for my current DVI displays.)

On the other hand, I too, like some other current Mac Pro owners, am worried about the lack of internal expandability. My current Mac Pro has four internal hard drive bays and two internal optical drive bays. And they are all full. I suppose I can live with external optical drives (although that will be yet another expense). But no internal bays for hard drives? Given that the new Mac Pro will come with an SSD, unless Apple has managed to perform an astounding miracle and make a 4 TB SSD that is actually affordable, there will obviously be a need for more storage, and that will mean that we’ll all be forced to purchase external Thunderbolt drives, again at additional cost.

I don’t have any problems with Apple adopting the new Thunderbolt 2 interface as long as it’s fully backward-compatible. But even Thunderbolt 1 offerings are still very limited today. It’s possible that, once the new Mac Pro has been on the market for a while, there will be a wider range of offerings at more affordable prices, but until then… In addition, if one of the key features of the new Mac Pro will be how quiet it is, how will this work in the real world if one is forced to have several external hard drives made by third party vendors who don’t necessarily put that much emphasis on noise reduction? (LaCie, I am looking at you.) It’s all well and good to have a quiet, superfast Mac Pro, but how quiet will the whole setup be once you’ve added a couple of external hard drives?

The current information provided by Apple is very short on details, technical specifications, etc. I guess we’ll have to be more patient. But it also looks like I might have to wait a while even after the new Mac Pro starts shipping, simply because the new computer is unlikely to be cheap and, with the additional expenses associated with the transitions (video adapters, external drives), it could turn out to be a very expensive proposition indeed.


Word 2011: Constant repagination while typing in footer

Posted by Pierre Igot in: Macintosh
June 7th, 2013 • 4:37 pm

If there is one thing about Microsoft Word for OS X that you can say, it is that it is never boring. For sheer entertainment value, no other piece of OS X software will ever give you a scene such as this one:

Here’s the context: I am working on a Word document created by someone else, which contains a mix of formatted text and various graphics. Nothing too stupendously complex, but not your average two-page letter either — which is apparently the only kind of document that Microsoft Word is actually able to handle competently (although even then…).

Then I go to the footer of the document (in page layout view, called “Print Layout” in Word), I select a phrase that I want to replace and I start typing “Version provisoire”…

As the movie clip above demonstrates, Word has somehow decided that, for this particular document, in the footer, it needs to repaginate the entire document each time I type another character… Of course, repagination is supposed to be a background task, but as you can see, when the document is a few dozen pages, even if the repagination is fast, it has a very real impact on the application’s responsiveness, to the point that Word eventually gives up and even ignores a few of the letters that I’ve typed (in the first word).

How we got to this point, I have absolutely no idea. But the very fact that this scenario is possible, that it indeed happened in front of my very eyes, as this recording proves, is rather representative of how broken Word 2011 is as a piece of OS X software.

Given that Microsoft clearly has no interest in fixing its software, I have absolutely no interest in solving this particular problem. All I can do is work around it and then try to forget that it ever happened… because, as far as entertainment goes, the experience of using Word 2011 for OS X can be pretty haunting. You just never know what might happen next. It is a proper thriller alright!


OS X 10.8.4: Yet another big disappointment

Posted by Pierre Igot in: Macintosh
June 4th, 2013 • 10:11 pm

What’s the point of submitting bug reports? That’s the question I used to ask myself when it came to Microsoft and its crappy software, but I must say I am increasingly asking myself the exact same thing with Apple and its own software offerings.

The two pieces of Apple software that I use the most are OS X itself (including all the free applications that are bundled with it, like Mail, Contacts, Safari, etc.) and the iWork suite of applications. I don’t need to remind anyone of the fact that there has not been a proper iWork upgrade in four years, and that even incremental updates have been very limited (mostly for things like iCloud compatibility), with no sign of Apple’s interest in fixing bugs or making improvements. It’s beyond embarrassing — but it’s a separate issue.

OS X itself is different of course. It’s the operating system, so Apple has no choice but to provide regular updates and upgrades. But increasingly, first when testing (via my AppleSeed membership) and then when using OS X updates, I find myself wondering, again and again: “Will they ever fix this? and this? and that?”

There are numerous annoyances in the operating system and the bundled applications that are simply not going away. And I am afraid I have to report that the latest OS X 10.8.4 update does nothing to address this situation, to the point that I really do wonder what it actually fixes. The update probably fixes some things for some people, but it puzzles me to no end that none of these things have anything to do with the bugs and flaws affect me in my use of OS X. I obviously don’t expect Apple to magically fix each and everyone of all the bugs that affect me in a single update, but at least some indication of incremental progress would be welcome. As it is, since the crucial improvements provided by OS X 10.8.2 (especially with the restoration of a “Save As…” behaviour that made sense), there has been… pretty much nothing.

On the contrary, Apple has actually introduced new bugs, some of which are very serious ones, and OS X 10.8.4 still does not fix them as far as I can tell.

The big one for me and many other Mac Pro users is that the kernel panics for Mac towers with dual GeForce video cards driving multiple monitors are still not gone.

As far as I can tell, Apple has been working on the problem. Throughout the testing phase for 10.8.4, “graphics drivers” was listed as what they call a “focus area”, i.e. something that was being worked on in the update. And indeed two of the three kernel extensions involved in the kernel panics, namely IOGraphicsFamily.kext and IONDRVSupport.kext, have been updated. (I also was contacted directly via AppleSeed about the kernel panic issue after they read my post on OS X 10.8.3. They gave me further instructions, and I followed these instructions and sent them the additional feedback they requested, but then I never heard from them again.)

My experience on my Mac Pro with two NVIDIA GeForce GT 120 cards driving two 30-inch Apple Cinema displays has been that, while the kernel panics are certainly far less frequent and less random than they were under OS X 10.8.3, they are still not entirely gone. I still had one a week ago while viewing a Flash video clip in Firefox on my secondary monitor. Several other users report that this situation is a particular trigger.

I view most of my on-line video in HTML5 and avoid Flash as much as I can (with the help of ClickToPlugin), but there are still situations where Flash is needed, and it’s massively disappointing that Apple still has not been able to fix this problem properly. (I also seem to be somewhat fortunate in that I do not need to use several other pieces of software that also seem to be triggers of the kernel panics.)

Lest you think that we are talking about a non-standard Mac Pro configuration here, I should reiterate here that the dual GT 120 cards were a standard customization option offered by Apple itself for the Mac Pro back in 2009. This is not a Mac Pro where I installed a third-party video card myself. This is a Mac Pro model configured and sold by Apple itself, which you would therefore expect them to continue to test on a regular basis with common configurations. While dual-monitor users like myself might be a minority, we still are a sizable group, and we also happen to have spent quite a chunk of money on Apple hardware.

Other Mac Pro users have given up on Apple ever fixing this problem and replaced one of their GT 120 cards with an ATI Radeon HD 5770 Graphics Upgrade Kit (sold by Apple via the Apple Store as well). This seems to be a permanent fix for the problem (if you’re willing to spend $249 + tax on it), which clearly demonstrates that the problem is with the NVIDIA drivers that Apple ships with OS X when running multiple monitors. (The cynical customer in me half-suspects that Apple’s apparent unwillingness to address this problem is in direct correlation with the number of ATI Radeon HD 5770 Graphics Upgrade Kits that they have been selling in the past few months…)

The bottom line here remains that Apple clearly knows that this has been a problem for nearly nine months now, but that they obviously don’t care enough about a small minority of professional users who happen to have purchased Mac Pros with multiple NVIDIA video cards in the past few years. I have said it before: I find this profoundly shocking. Even if Apple’s priorities are in the mobile market these days, this is inexcusable.

Just as inexcusable is the fact that one of Mountain Lion’s key features, namely the ability to restore applications to their last state after they are quit and relaunched or after the system is rebooted, still does not work properly. On my system at least, Safari itself still fails to restore the most recent session, instead restoring some unidentified previous session missing many of the tabs and windows I might have opened most recently. (And OS X also still fails to draw the expected drop shadows around Safari windows on the secondary screen, but here again, it’s only a problem for people with dual monitors, and possibly only a problem with NVIDIA cards, so I don’t expect Apple to fix that one any time soon, especially since it’s only cosmetic.)

OS X 10.8.4 also still often fails to restore windows in TextEdit, Preview and Numbers ’09 after a restart, at least on my system. Instead, it reopens the applications with no document windows whatsoever, and I’m left with having to navigate my “Open Recent” menus to try and find the documents that I need to reopen.

Restoring an application’s state is such a key feature in Lion and Mountain Lion that I simply can’t understand what is taking Apple so long to fix this problem — unless of course it just happens to be one of these bugs that only affects a minority of users and that Apple has not bothered to put much effort into trying to reproduce. Indeed, in my experience as a beta tester of Apple products, if you identify a bug that is not reproducible in an extremely straightforward manner, i.e. a bug that does not affect all users of OS X equally, the onus is apparently on you — the unpaid volunteer who is willing to experiment with beta software at his or her own expense — to provide Apple with all the required information to reproduce it on their machines. Except that even when you do provide as much information as you can, including all the steps you can think of, along with your detailed system profile, your installation logs, and so on, it’s still apparently not good enough for them, and I don’t really see any sign that they really do try to reproduce those problems that just happen to be a bit harder to reproduce.

Instead, you get… no feedback whatsoever, no sign that anyone is working on the bug, and then it does not get fixed in the next update and you cannot help but wonder: Should I submit a new bug report? Did it somehow fall through the cracks? These are all, to me, signs that the system for testing software and fixing bugs is indeed rather broken at Apple these days.

Other on-going issues in Safari include the dreaded “webpages are not responding” error message, which still occurs on a regular basis. It effectively forces you to reload all your currently open web pages, and is really just an application crash without the crashing part. One day, maybe, we’ll have a Safari web browser that has more than one “web content” processing thread and does not force you to reload all your pages when a single web site is causing the application to act up.

I’ve also noticed that, quite often, when Safari starts acting up, the responsiveness issues can get mixed with unpredictable cursor behaviours. For instance, I get the Spinning Beach Ball of Death, but Safari somehow still remains responsive and allows me to continue to switch windows and scroll through web pages — all that, while the SBBoD continues to spin. It’s quite puzzling and distracting, and the only remedy is usually a complete application relaunch… with its own associated problems, of course (see above!).

Mail is another OS X application that has numerous on-going problems that the 10.8.4 update fails to fix. It still does not allow me to remove attachments from my sent messages. Fortunately, someone has come up with a (weird) workaround that works for me and only requires one extra step, but still… Just how long is it going to take for Apple to acknowledge the problem and fix it? (Of course, it does not seem to affect people who only use an iCloud account in Mail. Just the rest of us with our own separate provider accounts.)

Mail still suffers from other problems introduced in Lion, nearly two years ago, such as the fact that the typing buffer ignores some keystrokes when you open a new message, that messages lose their rule-based colour when replied to, and that the blinking | cursor in a message that you are in the process of composing becomes invisible if you hide Mail while it is in the background and then return to it. (This bug also still affects other HTML-based applications, including the Robert dictionaries and the stand-alone web browsers created with Fluid.)

More recent problems that remain present in 10.8.4 include the fact that, on my system at least, there is often a delay between the time I click on the “Send” button and the time Mail actually closes the message window and proceeds to send the message. Sometimes this delay can be nearly 10 seconds, with no visual indication of what’s going on. Similarly, when I attach a document using the “Open File…” dialog box, once I have selected the file and closed the dialog, there is often a delay before the attachment is actually inserted in the message I am composing, and if I happen to move my cursor during that delay, the insertion position moves along with it and the attachment can get inserted right in the middle of a sentence, when it finally gets added to my message.

None of these issues are deal-breakers, but they are all constant annoyances, and they add up to a pretty frustrating experience. Most important, some of them have now been with us for a long time, sometimes years. Again, the question is: Exactly how long is it going to take for Apple to fix them?

Rather than gradual elimination of bugs over time, what I am observing these days is the persistence of old bugs and the regular addition of new bugs with new updates and upgrades. Again, I suspect that some things do get fixed for some people as part of these updates and upgrades, but for quite a while now, my personal experience has been that none of the issues that affect me ever get fixed and that, on the contrary, new issues are introduced on a regular basis.

Yes, in the big scheme of things, these are “details”. But there was a time when Apple’s legendary “attention to detail” actually manifested itself not just in new, amazing products, but also in incremental improvements of their existing products. In all honesty, I have seen very little of this in the past few years, to the point that I am starting to feel quite depressed about the whole situation. I simply don’t know what can be done about it, and exactly how many billions of dollars Apple needs to have in the bank before it finally starts spending some of that money on fixing problems.

While I fully respect Apple’s success on the mobile front, there is little denying that it has come at the expense of “serious” computer users whose life does not revolve around Facebook, Twitter, and Angry Birds. The very fact that we don’t have a release date for 10.9 and no details have even been announced yet, even though Apple promised a 1-year cycle last year, and the rumours that this has to do with OS X engineers being asked to focus on iOS instead, is just more confirmation that OS X simply is no longer a priority.

As a business, Apple has of course every right to focus on what drives its current success, but it makes one wonder where we’d be today if OS X had remained a bigger part of Apple’s overall activities. We’ll never know, but of course, since the alternatives (Windows? Linux?) are still more or less what they have always been, i.e. not really viable alternatives for demanding professionals, we’re effectively stuck and have to put up with what we have.


Swinsian: Fantastic iTunes replacement for music collectors

Posted by Pierre Igot in: iTunes, Music
May 7th, 2013 • 1:28 pm

As a music lover, I collect, among other things, lots of bootleg recordings of live Prince concerts. Back in the pre-Internet era, I actually spent some of my hard-earned cash on bootleg LPs, CDs, and VHS tapes. But thankfully, with the advent of the Internet, it has become possible to build a very decent collection of bootleg recordings without spending a cent. These recordings might not have the blessing of the artist himself, but at least there is no money involved and collecting them does not deprive the artist of any revenue. (My knowledge and appreciation of these bootleg recordings might affect my perception of the artist’s œuvre, but it certainly does not have an impact on my decisions when it comes to purchasing the artist’s official releases: I still buy everything he puts out.)

I also follow a number of other artists fairly closely, and also have a wide range of other musical interests.

This means, among other things, that I have amassed a rather vast collection of recordings, which needs to be maintained and managed. Most of it is now in digital form. (I still buy lots of music on CD, but I convert everything into digital files on my hard drive as well. The CDs are effectively my hard drive backup, and the medium for listening to the music on my main sound system, whereas I listen to the digital files with the sound system in my office.)

What are the options for managing these digital files? Well, on the Mac side, there seems to be pretty much only one option, which is iTunes. For years now, I have been praying for the introduction of some kind of version of iTunes optimized for music collectors — a kind of iTunes Pro, if you will. But of course, hoping that Apple itself would release such a product is nothing more than a pipe dream.

Instead, as time goes by, iTunes is becoming more and more bloated with things that are at best marginally useful to the music collector, and it’s becoming slower and slower, and buggier and buggier. Meanwhile, of course, my music collection is not getting any smaller.

A couple of months ago, I decided that I had finally had enough. While the iTunes 11.0.1 update released in late 2012 did address some of the worst issues introduced with iTunes 11, I still found myself constantly looking at the Spinning Beach Ball of Death, even for the most mundane of tasks (like starting or stopping playback!). In addition, knowing that this was at least partly due to iTunes constantly writing and rewriting the bloated files called “iTunes Library.itl” and “iTunes Library.xml” in the ~/Music/iTunes/ folder was not reassuring at all.

I had also started experiencing new bugs that were beyond irritating. For example, each time I imported music from an audio CD into my iTunes library, when I tried to eject the CD from within iTunes, the application would simply freeze altogether. (The only workaround was to quit iTunes after each import, eject the CD in the Finder, and then relaunch iTunes. Given that the very process of quitting and launching the iTunes application itself is slow as molasses, that was not really acceptable either.)

I was unable to reproduce the issue in a different user environment on my machine, so clearly it was a problem linked to my existing user environment and my existing iTunes music library. But I was not interested in exploring any of the more drastic troubleshooting options (like rebuilding my library from scratch, etc.) that would be required to try and identify the cause of the bug.

I then started exploring alternatives. First, I tried a couple of applications that were essentially alternate music players, like Ecoute or Fidelia, which would let me play the music from my iTunes library without having to launch iTunes itself. But I was not really satisfied and, besides, these solutions did not address my main concern, which was that even managing my music collection had become way too painful with iTunes. I needed a new music collection manager.

Then I stumbled upon Swinsian. I tried the demo and really liked what I saw. Unlike the other alternatives, it was an actual replacement for iTunes, not just as a music player, but as a music library manager. It could handle all the music file formats that iTunes supported (and then some). It offered tag editing, playlists, various view modes, etc.

But could it really replace iTunes? The first test was to see how it would handle my large collection of nearly 100,000 tracks. It passed that test with flying colours. Even the importing process was far shorter and smoother than I expected it to be. I made sure to configure it to only import the track info and not actually create duplicates of all my existing music files, and it was just fine with that.

Then I started playing around with the software, and began to feel increasingly confident about its ability to actually replace iTunes as my music library manager. It had some shortcomings, but it was very promising. I started sending reports about apparent bugs and issues to the developer, and was delighted to find that he was very responsive, even though I had not even purchased the software yet!

So I took the plunge, and decided to give it a real try. It wasn’t like I was taking a huge risk: If it didn’t work out, I could always go back to iTunes after a few weeks and simply update my library in iTunes to match the changes made in Swinsian. I would inevitably lose or have to redo some of my work in the process, but it would not be anything too catastrophic.

I purchased the software and set about using Swinsian in earnest for the various music library management tasks that my music collection usually requires. I started encountering other issues, of course, but was, once again, delighted to find that the developer would not only respond to my questions in a very positive and constructive manner, but would also actually implement bug fixes and enhancements in direct response to my comments, sometimes within a manner of days or even less! (I quickly opted to become a beta tester for new builds of the application, of course.)

This pattern continued for several weeks and it didn’t take me very long to decide that I had definitely made the right choice in switching from iTunes to Swinsian to manage my digital music collection. And now, as a long-time Mac user and iTunes user, I feel that it is almost my duty to report on what I have been able to experience and accomplish in Swinsian, with the help of its developer.

There is lots of material to cover and I won’t have time to mention everything that needs to be mentioned, but I’ll do my best to try and describe some of Swinsian’s key benefits and how it can be used as a replacement for iTunes for the purpose of managing and using one’s collection of digital music files. Everyone has different needs, of course, but I hope that sharing my experience will help others see that a switch from iTunes to Swinsian might be the best thing that they could do to regain control of their music library management experience.

Speed

This is definitely the main thing for me, and Swinsian really delivers. It is specifically designed for handling large libraries, and succeeds where iTunes fails miserably. It launches fast, it has a very responsive user interface, it lets you browse your large library quickly and smoothly, it lets you import tracks and edit tags efficiently, and it even has a Find/Replace feature that supports regular expressions, which is a crucial time-saver when you need to edit large numbers of tags in specific ways.

In fact, I am shocked to see how fast Swinsian is. Because of iTunes’s long-standing shortcomings and sluggishness, I had somehow managed to convince myself that the slowness was due to the large size of my library and to the intrinsic performance limitations associated with accessing and writing large numbers of files on a conventional hard drive. (My music collection is far too large to fit on an SSD drive.)

What Swinsian proves to me is that this was a false impression and that iTunes’s performance problems are not due to intrinsic limitations, but instead to the bad quality of iTunes itself as a piece of software. (Given Swinsian’s speed, the bottleneck is obviously not OS X itself or the file system.) I am more than surprised to see how fast Swinsian can, for example, modify the “Album” tag for dozens of tracks in a fraction of a second. I select the tracks, I edit the “Album” tag, I press Return to validate the changes — and my hard drive churns for about half a second, and that’s it! The same task in iTunes, at least for me with my large library, would be interrupted by multiple occurrences of the Spinning Beach Ball of Death and would easily, depending on the circumstances, take 10 or even 20 seconds sometimes.

It is really ridiculous how much faster Swinsian is, and it’s an eye-opener regarding the poor quality of iTunes itself as a piece of software, especially with large libraries.

Tags

With Swinsian, you can finally say goodbye to the semi-modal dialog box for editing track information. The information is displayed in a pane on the right-hand side that can stay visible at all times and is immediately updated based on the current selection in the main pane of the window. It takes a bit of time to adjust to this change (and the absence of visible field edges when the focus is not on the tag field is a bit disconcerting at first), but you quickly realize how pleasant it can be to have a non-modal interface for editing tags:

Swinsian-TrackInfo

Of course, the tags are not in the same order as in iTunes, so there is a period of adjustment there too. And Swinsian does not support as many tags as iTunes does. Notable absentees include iTunes’s sort fields. Swinsian only supports the standard “Artist” and “Album Artist” tags, so if you are used to sorting your artists by last name in iTunes, you will not have that option here. At first, I was a bit disappointed by this, especially given all the effort that I had put into tagging my tracks properly in iTunes in order to get, for example, all my Miles Davis music to show up under “D” rather than “M”.

But the truth is that, even after all these years, I was still very far from having properly tagged everything in my music library anyway, and of course, because of the limitations of the sort field feature in iTunes itself, it was a never-ending task, because there was nothing smart about this feature, and every time I imported a new Miles Davis album in iTunes, for instance, I still had to manually tag it properly in order to avoid having it show up under “M” rather than “D”. (A really smart sort field feature would have rules that would automatically add the right tags, at least for known artists with known sort fields, but also for obvious names. I mean, how hard can it be for a computer program to figure out automatically that, in “Neil Young”, “Young” is the last name and “Neil” is the first name?)

So basically I had to abandon the idea of having a music library with alphabetically-correct sorting. It’s not the end of the world for me, because I do a large part of my library browsing either via playlists or with the search feature anyway. When you access your Miles Davis tracks mostly by selecting a playlist or by typing “Miles Davis” in the search field (with instantaneous results in Swinsian, thank you very much), it does not much matter whether Miles Davis shows up under “D” or under “M” in alphabetical lists of artists.

One thing that the Swinsian developer has implemented is ignoring “The” when sorting things alphabetically. Initially, this only worked for artists, but he has recently added support for this same feature in track titles, so that “The End” shows up under “E” and not under “T” in alphabetical order. (And you can also browse lists by typing, which also correctly ignore the “The”. Ignoring “A/An” is not supported at this time.)

When editing tags, it is also now possible to use the familiar command-P and command-N shortcuts (as in “Previous” and “Next”) to jump from track to track in the Track Info pane without having to change the selection in the main pane. Even though the Track Info pane is not modal, it would still be a pain to constantly have to grab the mouse to select the track(s) whose info you want to modify.

In addition, Swinsian is a well-behaved OS X application, with support for the built-in spell checker, substitutions, etc. And it also provides text transformation commands for changing text to lowercase, title case or all caps.

Last but not least, Swinsian has a very powerful Find/Replace feature (separate from the regular search field) that supports regular expressions. This to me is pretty close to the Holy Grail of tag editing. Regular expressions have a bit of a learning curve, but they are such a powerful tool for editing large numbers of tags that I don’t think any owner of a large music collection can really do without them. (I know I did until now, but really I was just waiting for such a feature to finally become available.)

You can use the Find/Replace dialog without learning about regular expressions, of course, and it already enables you to execute batch operations that are impossible in iTunes, but regular expressions are where this feature truly shines.

In the process of collecting bootleg Prince recordings, for example, I often get, among other problems, tracks where the “Title” (song title) field contains the song title prefixed by a track number (as in “04 – Purple Rain”). Since there is already a separate track number tag, this is highly undesirable and the track number prefix needs to be removed. But how do you remove it? Well, without regular expressions, you have no choice but to do it manually, one track at a time. (Experienced iTunes users know that there are AppleScript scripts for iTunes for this type of thing, but they too are affected by the major performance issues in iTunes.) In Swinsian, you can just use a regular expression like:

[0-9]+ - 

(which means “any sequence of one or more digits followed by a space, a dash, and another space”) in the “Find:” field and replace all occurrences with nothing. While you cannot restrict the Find/Replace operation to the current selection of tracks, you can restrict the operation to tracks added recently only (in the last X days). And you can also preview the changes to make sure that the operation won’t affect any tracks that you don’t want it to affect.

With regular expressions, the possibilities are endless. You can clean up hundreds, thousands or tens of thousands of tracks with lightning-fast batch operations. It is an essential tool for managing a large collection of digital music files.

Controls

If you are used to using your keyboard’s media keys to control playback in iTunes, the same media keys will work Swinsian. On my aluminium Apple keyboard, for example, I can use fn-F8 from anywhere in OS X to play/pause playback in Swinsian, and fn-F7 and fn-F9 to jump to the previous and next tracks. (You don’t need the fn modifier key if you use OS X’s default keyboard settings for the aluminium keyboard.)

Controlling Swinsian playback volume specifically (without changing the overall volume level for the system) from anywhere in OS X is a bit more complicated, but one other key aspect of Swinsian is that it’s scriptable.

For increasing the playback volume, I ended up using Keyboard Maestro to define a macro that runs this AppleScript script:

tell application "Swinsian"
	set myVolume to volume
	set the volume to (myVolume + 0.05)
end tell

and displays a quick notification in Mountain Lion’s Notification Center, and I assigned the global cmd-shift-F12 shortcut to it. I have a similar macro for decreasing the playback volume in Swinsian.

I am still in the process of experimenting with other AppleScript scripts. For example, for my Prince bootleg recordings, I also need a script to “albumize” a selection of tracks, i.e. to automatically number these tracks in the right order in the “Track Number” field. (Typically, tracks that come with a track number prefix in the “Title” field also fail to include a track number in the “Track Number” field.) Even though my scripting skills are quite limited, I already have a script that appears to be working fine:

tell application "Swinsian"
	set p to selection of window 1
	if p is not {} then
		set c to count of p's items
		repeat with i from 1 to c
			set t to item i of p
			set the track number of t to i
		end repeat
	end if
end tell

I initially had some performance issues with this script. But again, the developer has been very prompt to resolve them.

Interface

If you are one of the many iTunes users who were dismayed by the fact that Apple removed the small pane displaying the album art in the main window in iTunes 11 in list view, I am pleased to report that the track information pane in Swinsian includes such a section.

Sometimes, Apple’s design decisions are maddeningly stupid, and at some point you find yourself wanting to throw your arms in the air and give up on the whole thing. Well, with Swinsian, you can give up on iTunes and get a proper list view mode with album art again:

Swinsian-AlbumArt

Swinsian also brings back the option to have a browser view with columns on the left — another key feature that was removed by Apple in iTunes 11. And it’s quite flexible:

Swinsian-BrowserOptions

In many respects, Swinsian offers a user interface that is what the iTunes interface for music collectors should be, and no longer is.

The playlist management features are similar to the ones in iTunes. Due to restrictions imposed by Apple itself, Swinsian is unable to import smart playlists from iTunes, and imports them as regular (static) playlists instead. But you can easily recreate your favourite smart playlists in Swinsian itself.

The one big thing that is still missing in Swinsian, in my view, is the ability to open playlists in separate windows. It’s another thing that Apple removed in iTunes 11, and I was hoping to get it back with Swinsian. But the developer tells me that he’s thinking about it, so I am quite optimistic.

iTunes and Swinsian Cohabitation

Even if, like me, you decide to embrace Swinsian as your primary music library manager, it’s unlikely that you’ll be able to get rid of iTunes altogether. For one thing, you might be a regular iTunes Store customer and want to continue to be able to purchase music from the store.

The good news is that you can have both. All you need to do, after importing your existing iTunes music library into Swinsian, is to use Swinsian’s “Watched Folders” feature to keep an eye on your iTunes Music folder automatically:

Swinsian-WatchedFolders

On my machine, iTunes is configured to save new music files to a folder called “Music 2009” (where I consolidated my entire music library back in 2009, hence the name). Now that I’ve moved to Swinsian as my main music library manager, I have changed the iTunes setting so that music files are no longer copied to this folder when I add them to the iTunes library (so that I don’t end up with all kinds of duplicates when I add stuff from my Swinsian library to my iTunes library). But when I buy some music from the iTunes Store in iTunes, it still gets added to my “Music 2009” folder. And now, with Swinsian’s “Watched Folders” feature, these newly-purchased tracks are also automatically added to my Swinsian library as well. (Again, to avoid multiple copies, I have configured the “Watched Folders” settings for my “Music 2009” folder in Swinsian to “Just Add Tracks”, as opposed to copying them.

This way, when I import new music (from a ripped CD, for example), it gets copied to my Swinsian library folder. But when I purchase something from the iTunes Store in iTunes (which gets copied to my iTunes music folder), it simply gets added to my Swinsian library, without getting copied to my Swinsian library folder. This means that I have music files in two different places, yes, but both those locations are on the same hard drive, and I back up everything on that hard drive nightly.

The other important issue regarding iTunes and Swinsian cohabitation is managing the music that’s stored on iOS devices. Initially, I thought I’d have to try and keep my Swinsian library and my iTunes library in complete sync. But I’ve since discovered how, once you’ve plugged in your iOS device and it shows up in iTunes, you can switch to the device’s “Music” pane and then drag and drop music files from Swinsian directly onto that pane in iTunes so that it gets added to the device, bypassing the iTunes library itself altogether.

(You can also easily select a bunch of tracks in Swinsian and simply drag-and-drop them into iTunes to add them to the iTunes library. As long as you have the right settings in iTunes, iTunes will simply add references to the tracks to its library and not copy the tracks. And it’s actually reasonably fast — as fast as iTunes can be these days, anyway.)

Of course, this will never be as good as complete iOS device integration or direct access to the iTunes Store, which only iTunes can offer. But it’s good enough for me and I suspect it might be good enough for many other music collectors.

I should also note that Swinsian includes support for AirTunes, which means that you don’t necessarily have to rely on iTunes to play the contents of your music library on your living room’s sound system, for example.

Importing CDs

Swinsian does not include its own feature for importing CDs. Instead, you can use a third-party application, such as XLD. In order to make the process smoother, you can use Swinsian’s “Watched Folders” feature to keep an eye on the folder where XLD saves the imported tracks, so that they will be automatically moved from that folder to Swinsian’s library folder and added to the library. (See the “Watched Folders” screen shot above.)

XLD is donationware and is a pretty powerful tool, which lets you use alternate sources for track information and album artwork. And it does not freeze when you try to eject a CD after importing it!

Conclusion

There are several other aspects that I have not mentioned about Swinsian (podcasts, shuffle, last.fm support, etc.), simply because I don’t really use them myself, or because I am running out of time and space. But I hope I have given you a good idea of what Swinsian can do as a replacement for iTunes.

Of course, by switching to such a tool for managing my music collection, I am also taking a bit of a long-term risk in that I am now dependent on a single developer continuing to develop, update, upgrade, and otherwise improve his software. With the passing of Evan Gross, developer of Spell Catcher, less than a year ago, I am more than sufficiently aware of the risks associated with excessive reliance on independent developers.

But at this point in time, there is simply no reason to believe that Apple will ever make iTunes a user-friendly tool for music collectors again, especially those with large music collections. So it is, once again, a risk that I am willing to take. One thing to keep in mind is that the work you might do on tagging your music collection properly, for example, is not application-dependent, since the tags are stored with the music files themselves, which remain fully accessible and transferable.

I certainly hope that Swinsian’s developer will continue to develop his application for many years to come, while preserving its essential qualities of speed, responsiveness, and power — qualities that are sorely lacking in iTunes today and show no sign of being part of Apple’s priorities these days. Also, if I can convince other people to switch to Swinsian, this will of course also increase the likelihood that it will continue to be developed and improved, so I am not being completely selfless here!

If you are the owner of a large music collection and are endlessly frustrated by iTunes’s numerous flaws and major performance issues, I strongly urge you to give Swinsian a try. It’s not a flashy piece of software that tries to reinvent the wheel with custom UI controls or crumbles under its own feature bloat. It’s an extremely solid, super-fast, standards-compliant application that really gets the job done.


Word 2011: Header formatting madness (part III)

Posted by Pierre Igot in: Microsoft
May 5th, 2013 • 8:30 am

I have received more feedback about the issue with character-level formatting in the header of a Word document that I was working on the other day.

Yesterday, I indicated that the problem had to do with the fact that the text in the header, even though it didn’t look any different from regular text, was actually a reference to metadata associated with the Word document.

Since then, a couple of readers have written to confirm that, in Word 2010 under Windows, there is a visual difference in the header: “The title is shown within a light blue rectangle with the tag ‘title’.” I obviously cannot reproduce this on the Mac and I don’t have a copy of Word 2010 for Windows to confirm this myself. All I can say is that, in Word 2011, there is no visual difference whatsoever between this header text and regular header text.

Apparently, this all has to do with a feature called “content controls” that was introduced in Word 2007 for Windows. Evidently, this feature was never introduced in the Mac version of Word, but Word for OS X has to be able to open and display Word documents created in Windows. Since these Word documents can contain such “content controls” but Word for OS X does not have a “content controls” feature, what does the MacBU do? It just pretends that the feature does not exist and displays content controls as regular text.

But of course if you start interacting with such text, you end up encountering issues such as the one I experienced the other day. What I had in my document was apparently what is called, on the Windows side, a “plain text content control”:

… if you italicize one word of a sentence that is in a plain text control, all the text inside the control is italicized.

There is such a thing as a “rich text content control”, which can contain formatted text, but of course even if I had wanted to replace the plain text control with a rich text content control in my header, I wouldn’t have been able to do so, since Word for OS X does not have a “content controls” feature.

I still have no idea whether the author of the Word document that I was working on had even intended to use this “content controls” feature himself or it was used accidentally or unintentionally in the process of creating the document (maybe through the use of an existing document template or because Word has some other automatic behaviour that inserts such things without the user really understanding what’s going on).

What seems pretty clear to me is that these “content controls” are primarily intended as some kind of replacement for form fields (which are supported in Word for OS X). But it’s far from obvious to me that there is any benefit to insert a reference to the “title” field in the Word document metadata as opposed to simply typing out the title, whereas there are obvious drawbacks, such as the fact that the “title” field in the metadata only contains plain text.

You also won’t be surprised to hear that, based on the article about content controls mentioned above, the feature is at best half-baked even on the Windows side:

… despite their enormous potential, Microsoft has failed miserably in resolving long standing bugs and in providing enhancements to functionality to fully realize this potential.

What else is new?

What isn’t new either is the fact that Mac Word users are treated as second-class citizens and left to deal with mysterious, unexplainable behaviours all on their own.

Thanks to L. H. and Dan for their additional feedback.