Microsoft’s Rick Schaut on Word 6
Posted by Pierre Igot in: MacintoshFebruary 27th, 2004 • 5:23 am
Microsoft developer Rick Schaut has an interesting blog entry on the Macintosh software fiasco known as Word 6.0 for Macintosh. It is written from a developer’s point of view, so it’s pretty hard-core, but it’s interesting nonetheless.
Also interesting are the comments added by readers, and Rick Schaut’s reply to these comments. For example, to a comment from Sandy McMurray about whether the next version of Office for Macintosh (Office 2004) would support Mac OS X’s built-in spell checker and Address Book functionality, here’s Rick’s answer:
Sandy, to be perfectly honest, it would be a serious amount of work to try to use the system spelling checker in Word. One reason is that Word’s spell checking is tied in with grammar checking. However, there might be a way that we can grab the list of custom words from the system dictionaries. I’ll pass the idea on to our proofing tools people.
Well, that says it all, doesn’t it? Why should the user care whether Word’s spell checker is “tied in with grammar checking”? Most of the people I know never use a grammar checker in the first place (not to mention Microsoft’s version of one), simply because the grammar of natural language is far more complex than any software currently available can handle. This means that most of the grammar checking done by tools such as Word’s grammar checker is useless, because it flags as wrong things that are perfectly right, and doesn’t catch obvious grammar errors because it’s not smart enough.
And the fact that the idea of supporting Mac OS X’s built-in spell checker still needs to be passed on to Microsoft’s “proofing tools people” at this particular stage in the development of Office 2004 is, well, to put it simply, ridiculous. Mac OS X has been out for years, and its built-in spell checker has been accessible to Carbon applications for a while too. Do Microsoft people really need to be told at this point in time that Mac OS X users would like Word to support Mac OS X’s built-in spell checker?
I’ve said it before and I’ll say it again: If you want a system-wide spell checker solution, check out Spell Catcher X. It works in all applications, including Microsoft Word and Excel, and it is also accessible through Mac OS X’s built-in spell checker.
As for Office 2004 supporting Address Book, well, Rick Schaut doesn’t even answer that question. That should tell you how much of a priority it is for Microsoft. (Remember that it’s not in Microsoft’s economic interests to support standards or Mac OS X’s built-in features, unless they can “customize” them in a way to lets them pretend that they support them while they actually don’t.)
I suppose that, in 2011, we’ll be reading the blog entries of someone working for Microsoft who will reminisce about how Microsoft finally understood what it meant to be “Mac-like” and user-friendly after the disappointing sales of Microsoft Office X in 2003. Some things never change…
March 10th, 2004 at Mar 10, 04 | 7:52 am
Pierre,
The whole question of customer contact with respect to bugs is a huge problem. I’ll talk a bit about that too.
For now, though, the most effective place for users to go with issues is the appropriate microsoft newsgroup.
Yeah, I now that’s lame. There’s a lot we have to work out before we come up with something less lame.
Rick
March 10th, 2004 at Mar 10, 04 | 8:05 am
Rick:
Thanks for your clarification. I’ll be interested to see what you have to say.
Abstracting from Pierre’s latest post: There are two vital elements, one that some interest be shown in receiving bug reports –a few mac-specific customizations would go a long way– and the other than some kind of acknowledgement occur at least some of time.
Of course it would be even better to work _with_ knowledgeable users. Think of Pierre posting to the newsgroup comments to the effect that he’d been able to resolve a long-standing issue by working with you or your colleagues!
I would like to observe that MacOS X has greatly improved, user-visible facilities for debugging over previous versions of the OS and –to my limited knowledge– over most versions of Windows. (I guess XP “may-I-phone-home” bug reporting probably has similar capabilities, but the detailed data doesn’t seem to be available to users.) To me lack of use of these new capabilities on the part of MS is ample justification for speculation.
Thanks,
Henry
March 19th, 2004 at Mar 19, 04 | 10:47 pm
A very interesting discussion.
I switched to Mac from WinTel about a year ago, when I got sick of the constant operating system crashes (experienced on several computers over many years, so it wasn’t “my particular system”). Unfortunately, WordPerfect, which I consider a superior word processor, is long out of development on Mac. So I had to switch to MS Word. Yes, my publisher demands files in .doc format.
I spent a good part of the next six months converting my library of WP reference files into Word format. These were moderately formatted–a lot of hanging indents, some automatic numbering, various fonts, little else. It was a heck of a job. Hanging indents were a mess. Automatic numbers were gone. Fonts were consistently (and irrationally changed). This was the case, no matter whose conversion engine I used (Corel, MS, or third-party).
Now it’s all done, except for minor mistakes. Am I going to convert all these files to some other format in the future? Not likely. They tell me Mellel is great, but it has trouble with MS compatibility. Obviously creating a usuable .doc format is not as easy as Rick claims. I’m stuck with Word.
As for the “pause in entering text bug,” it happens all the time, on practically any document. Sometimes I’ll type something, and the last letter will not appear until I hit another key.
March 20th, 2004 at Mar 20, 04 | 2:33 am
William Shepherd:
Thanks for your addition.
I have an XP system literally next to my Mac. Both (historically, at least) have the latest version of Office, and I maintain both systems quite carefully. I don’t use the XP system much, so it is relatively close to a “stock” system.
Why duplicate? Because I had a client which specified documents in Word format, and which I knew uses PCs exclusively. The last thing I needed was to deliver a document that had messed-up formatting when read with Word XP. So I can compare … side-by-side.
Yes, relatively simple structures transferred over reasonably well. Anything beyond that.. well, there were always glitches. Since I avoid using any Word feature I don’t absolutely need for fear of crashes and/or document corruption, I can’t say how advanced things like auto-numbering transferred.
Fonts, and thus some page layout, didn’t always work very well. Now, this may in fact be due to some fundamental differences between fonts on the Mac and the PC, and I’ll allow for that. To a point. Does MS advertise cross-platform compatibility, or doesn’t it? Either it works or it doesn’t.
(I have to throw this in: I wanted to submit my material in FrameMaker, because my client had a documentation department on Macs using FM, and I had confidence in my ability to deliver near-final manuscripts that would render correctly after delivery. At some point, the client laid off all those folks, got rid of the Macs, and made no plans for supporting that function beyond having Word on the PC of every remaining employee. Eventually I was told that I would no longer be given writing assignments, not because my work wasn’t high-quality, but because they weren’t ever able to publish it.)
Having latest copies of Word on two different platforms has given me some interesting opportunities. For example, somewhere I have a Word file from my Mac that will no longer read correctly on the Mac, but it reads and looks fine on Word XP. Huh? If that doesn’t indicate inherent problems within MS’s formats and readers, I don’t know what would. If MS can’t provide reliable cross-platform transfers, I would not expect any kind of success translating from completely different formats, sorry to say.
Mr. Schaut asserts that the Word file format is public. Mr. Igot respond that the cited formats are for Office 97 — yes, the link Mr. Schaut gave is to an article dated 1997. Mr. Schaut replies that Word 97 formats are sufficient for “A huge market…” Huh? What year is this? Has the format changed since Word 97 or hasn’t it? Are the changes orthogonal or not?
As I said in an earlier post, I’ve done surveys to see if I could find either specifications or actual code embodying current Word file formats. I even downloaded the source for OpenOffice (or the its cousin, I forget) — an astoundingly large amount of code, which overwhelmed my source browser. All in all, the quarry is quite elusive. I’ve never succeeded in finding anything that is both current and comprehensive. Now that Google has significantly increased their page count, maybe I’d better try it again.
What I have discovered thus far is that the file format engineering design may not be either robust or easily transferrable. If I’m correct –and I may not be– you _must_ use MS development tools to reproduce the “container within container” object model they use. Again, I’m not sure, but I get the strong feeling that any given “thing” (maybe a paragraph) at some level of containment is vulnerable to corruption at some higher level (maybe a section). This is at least consistent with the problems of document corruption I’ve observed, though it is by no means proof.
By the way, I’ve _never_ observed the “pause in entering text” bug.
Thanks,
Henry
March 20th, 2004 at Mar 20, 04 | 4:25 am
Henry: In Mac OS X, you can actually use PC TrueType fonts directly without any kind of conversion. PC TrueType fonts usually have the “.ttf” extension. They work in OS X as well. I’d be curious to know if you’ve observed problems between Mac and PC with documents that only use such “.ttf” fonts.
Thanks for your FrameMaker story. I guess it’s a perfect example of how Mac users are forced to use Microsoft Word.
As for the file format, my guess is that, if we had a really free market where products can compete with each other in a healthy fashion, and if the Word format was truly available to any developer without requiring reverse engineering or huge development costs, then surely by now there would be a number of decent competitors for Word X under OS X. A simple example is the use of “frames” in Word — this pseudo-page layout tool that lets you put text or pictures anywhere on a page. I constantly receive 1- or 2-page documents that contain at least one such frame that was used to place some kind of logo or header in the document. Even Word’s own default behaviour creates such a frame in the footer of a document when using the Insert Page Number command, which many people are likely to use without noticing (that it inserts a frame instead of a simple page number text string). Is it easy for third-party developers to write a word processor that is compatible with Word documents that contain frames? I doubt it… Yet I keep receiving documents with such frames, because Word inserts them automatically in certain cases or people think that they can use Word for page layout purposes. Same thing for automatic bullets/numbering. Since it’s on by default and takes over when a user starts typing a list, people use it even without knowing.
The end result is that even a single page document created by an ordinary Word user with no advanced knowledge of the product can end up being a very complex file that contains tables, frames, field codes, automating numbering, and what not.
March 20th, 2004 at Mar 20, 04 | 5:27 am
William and Henry: Obviously not everyone experiences the stalls or notices them. But surely there are enough people out there who do experience them and are annoyed by them. This bug is far too obvious, even on my fast system. It is simply shameful that Word cannot perform the most fundamental task in word processing properly, i.e. typing text. Yes, I am a fast typist, so what? Is it illegal? Is it too much to ask that a dual GHz computer be at least able to keep up with my typing, especially considering that, in most circumstances, my CPU usage stays at approximately 30% on each CPU unit? All other programs are able to keep up with my typing just fine. Why not Word X?
March 20th, 2004 at Mar 20, 04 | 7:46 am
Pierre:
Thanks for your comments:
No, I haven’t tried using True Type fonts for portability, but I will, next time I need this capability. It seems to me that an application that promises cross-platform portability ought to provide at least some framework of support. Hmmm, what about something like this on Word XP: “This document was created on a Mac, with the following non-TrueType fonts. You’ll likely get better results –more equivalent look and placement– if you replace these fonts before transferring the file to the PC platform.” And vice-versa.
About frames, etc.: If I recall correctly, one of our favorite MVPs has strongly and publicly discouraged us from using them, an admonition I appreciate greatly –it confirms my own impression– and I’m glad to heed. If I recall correctly, frames don’t even work reliably with respect to to one file for one user on one platform. I think there’s also a rather complete and very negative appraisal of Word’s auto-numbering, possibly from the same source.
About file formats: Thanks for pointing out that the most trivial document may invisibily contain rather complex structures. From what I’ve seen by observing file sizes and looking at hex–ascii file dumps, you are correct. Now, there’s no sin in preparing a sophisticated structure when a Word file is created. After all, Word can’t know if the file will end up being one line or 1000 pages long. Equally, there’s no inherent problem in using a complex mechanism to support a convenience like autonumbering. Well, there is _one_ problem: It seems you and I all too often discover that such mechanisms are non-robust, never mind that we sometimes having trouble controlling them. I wonder: it is just us who experience this instability, loss of formatting ranging to unrecoverable file corruption? I would certainly like to hear what other users experience. It seems unlikely that it is just you and me.
About typing stalls: I’ve been a Unix user on and off for maybe 25 years. With respect to the Unix command line, I’ve observed a kind of natural rhythm, which includes some stalls. This has continued on the MacOS X command line and I’m OK with it. That said, if you run 10 major GUI apps, 9 of them don’t show this behavior, and one does, it is very reasonable to finger that app. Pure speculation: there’s something unfortunate about Word’s design and/or implementation history that requires a task to be done between characters. No harm done on some systems, and serious stalls on others. Hmmm, want to spend some time comparing system configurations? Maybe we could narrow this down.
Thanks,
Henry
March 20th, 2004 at Mar 20, 04 | 8:18 am
Henry:
The fondamental problem with frames, autonumbering and all these other things that do not work reliably enough and dramatically increase the complexity of the document is that, due to Word’s default behaviours, Word users who didn’t turn various default options off end up using them without knowing it! That is why Rick cannot claim that 1-page or 2-page documents should be easy enough for compatibility purposes with other applications. The frame around the page number is a typical example. If you are in Word X with a new blank document and you go to the “Insert” menu and choose “Page Number…” and just click on the “OK” button, Word inserts a page number in the footer of your document and that page number is… inside a frame! That’s the default behaviour!
As for the stalls, the problem with your hypothesis is that, as far as I can tell, Word is doing absolutely nothing during those stalls. It’s not doing anything else. The simple proof of this is that you can interrupt a stall by moving your mouse pointer. (That’s precisely how Evan Gross was able to work around the problem in Spell Catcher X 10.1. He uses sub-pixel mouse movements generated by his software during the typing process which interrupt the stalls when they occur and have no incidence on the actual position of the mouse pointer.) You can also interrupt the stall, like William says, by hitting another key on the keyboard. (I often use the Left or Right cursor key to interrupt a stall – although thanks to Spell Catcher X 10.1 I rarely have to do that anymore.)
If you can interrupt the stall without any consequence, that’s clearly proof that the stall doesn’t correspond to any essential activity that Word cannot do later. (The other indication that nothing is going on is that there is no change in CPU activity and no hard drive activity during a stall.) That would also explain why Word is the only application exhibiting the behaviour. (It’s a flaw in Word, not in the Unix architecture of Mac OS X.)
March 20th, 2004 at Mar 20, 04 | 9:41 am
Pierre:
I was always uneasy about that “Insert … Page Number…” anyway. It seems, well, uncoordinated with the explicit commands. (I need to check my records. I think I heard some helpful information about the relationship between these two a while back.)
As for stalls, I’m still stuck on the fact that I don’t experience them.
Wild speculation: Let’s say Word has some (obscure) reason to check that the machine’s network interface is “connected” between keystrokes. In my machine, with DSL, that’s not a problem. On yours, with dial-up, there’s a different mechanism and –as a result– different timing. There may also be further variation depending on dial-up or related configuration.
If Word does nothing visibly different whether the answer is “yes”, “no,” or if the inquiry is aborted by user activity, that would be consistent with what we all observe. (Right?) It’s a bug from your point of view and invisible to me. I’m not surprised that there’s no visible difference, either at the level of Word or in overall CPU activity — a inquiry of this type would simply be too small to register in any case.
It’s moderately believable that MS hasn’t been able to reproduce this issue, but there are, unfortunately, less innocent alternative explanations. Without further information…
…And further information would include data on the prevalence of reports of keyboard stalls, especially, if they appear to be relatively configuration-independent.
If Mr. Schaut is listening, I would request that he consider his situation if he personally experienced this issue. Please stand aside from Pierre’s quite justifiable expressions of frustration and share with us what you would do.
thanks,
Henry
March 21st, 2004 at Mar 21, 04 | 7:56 am
I really doubt that Word keep trying to connect to the Internet at regular intervals. It’s crazy, but not that crazy :). (On the other hand, I keep seeing these “Connecting to the printer” messages in the status bar when saving documents, so who knows…)
It would indeed be interesting to do a detailed comparison of the configurations of the various people who do experience the stalls, in order to isolate a possible source. But that’s Microsoft’s job to do, not mine!
It should also be noted that the stalls affect not just typing in Word, but also various other elements in the interface. For example, often when using the Find/Replace dialog, there is a stall after clicking on one of the button before the corresponding action takes place. Here again, jigging the mouse a bit interrupts the stall and causes Word to proceed.
March 21st, 2004 at Mar 21, 04 | 8:23 am
Pierre:
OK, the internet, a printer, the mothership, whatever. I simply wanted to provide a plausible model for what’s possibly going on — a process that doesn’t actually do anything other than record the status of some external device or process.
Could you summarize your success thus far at getting attention from MS, the MVPs, or Mr. Schaut on this particular issue? Any acknowledgement? Have you tried the Apple discussion boards? Do you have any ideas about how many people are observing keybard hesitations in Word? (Sorry, I just haven’t followed this issue.)
Your comments jiggle my brain a little bit about the Mac human interface. Most apps allow the MacOS to draw standard dialog boxes — it’s a look-and-feel consistency issue — and let MacOS handle keyboard+mouse input for these. I _think_ almost everything you ask the OS to do is capable of “calling back” to application-supplied routines to take care of the details. Thus, it’s possible that a dialog drawn by the OS could depend on console I/O routines supplied by Word. If so, there may be a way of asking the OS to tell us more about what it is being hot-wired to do. On the other hand, if Word draws its own dialogs it may also simply supply most of the underlying keyboard and mouse handling, down to the virtual ROM toolbox. In that case, it still might be possible.
I get a strong feeling that what we’re seeing here is _something_ built in to Word a long tim ago that has been carried forward over the years and is now biting some people. The original reason –if anyone remembers it– for the mechanism may have even been some limitation of the MacOS (rather “System”) of the day.
Thanks,
Henry
March 27th, 2004 at Mar 27, 04 | 4:21 am
I found your blog by googling “Word X sluggish” and have read your posts on the Word stalls. Just wanted to let you know that I’ve experienced the same exact problem since I installed Word X on my iMac G4/1.25 with 768 MB RAM. Extremely frustrating. Significantly degrades my productivity. Looking for light at the end of the tunnel….
December 7th, 2004 at Dec 07, 04 | 6:19 am
The Microsoft view is entirely pragmatic: They try a method. If it works — possibly it is sufficient if no important customers complain– it _is_ right.
The alternative view is that there exists some kind of objective standard of “rightness” and that a single individual can judge an implementation against that standard.
These discussions with Mr. Schaut get nowhere because of vastly differing assumptions.
December 6th, 2004 at Dec 06, 04 | 11:03 am
Rick said: “the only thing absurd in this entire discussion is your claim, based on your own parochial views and not on an objective survey of the needs of all Office users, that ?Microsoft is driven by interests other than usability and user-friendliness.?
Hmmmmmmmmmmmm … I was looking for information on a MSFT Word bug that has been plaguing Windows users since Word 6 was a pup. It’s STILL plaguing me in Office XP, so there are at least two users with the “parochial” view that Microsoft doesn’t care about the users. They don’t care about devellopers either – I just tried using Word’s filtered HTML with the MSFT HTML Help SDK and it borked on me. Their tools can’t eat their own dogfood!
I think the light at the end of the tunnel is found at openoffice.org because it does as good a job of reading MSFT Office as any one version of MSFT Word does at reading any other version (if it can read them at all). And there is a MacX version :)