norman@sdics.ucsd.EDU (Donald A. Norman) (05/25/87)
References: Summary: A short story of my tragedy and partial recovery. Then questions about how Word 3.0 stores things, and why. And listing of 2 other problems, while I am at it. For Word Gurus, plus the curious. So, I go ahead and use Word 3.0 to do the book. Glutton for punishment me, but I am addicted to the features. So there I am, four hours since the last incremental hard-disk dump, writing the chapter on system failures when: A routine back-up save to the hard disk hangs up. 66% written and nothing is happening. The 66% figure stays there, the clockface cursor sits there, but no disk noises. j-clock keeps running. 20 minutes later it is still there. So I take a few more deep breaths and reboot. [interactions? j-clock. Superspool Imagewriter spooler. Standard Mac +. DataFrame 20 Mbyte hard disk. Nothing else that should cause problems.] My file is there, 59K long, but when I load it, Word tells me it has only 839 characters and, worse, what is in it is my "todo" notes which are supposed to be in a different file, in a different subdirectory. There is a Word Temp 44 file some 77K long. Word can't load it. I poke around the files with Mactools and can't make any sense of what is there. Lots of text in both the trashed text file and the temp file, but it seems to be randomly taken from stuff written over the past 6 months. Stray stuff on the disk, one would think. In desperation, I change the TYPE and CREATOR on the Word Temp file to be WDBN and MSWD and just try loading the temp file into Word 3. It works! I only lose 2 hours of stuff (I back up every few minutes -- why did I lose 2 hours?) ---- QUESTIONS and OTHER PROBLEMS What is the purpose of the temp file? Why is it a backup? If it really is a backup, why doesn't MS word provide a recovery process for it. Anybody ever recover lost stuff before? Why can't I piece together the text in the trashed main file: why is it 59K long, with only 839 bytes of recoverable stuff -- the wrong stuff. The Table of Contents suddenly stopped working a week ago. That is, it works fine on short files, bombs, consistently on long files. It used to be an essential feature. Now it is useless. It seemed to bomb on the second pass, after pagination, during the collection of terms, and only when it got to the very last page of the manuscript. So, I try a GOTO page 99 (a non-existent page): bomb. Consistently. Now what. Anyone else have this problem? And every so often Word reverts to innocence. I boot the system and Word displays short menus. No specially tailored menus. No memory of all my styles. The Word Settings File is clobbered. Why? No clues. An incremental restore of the file saves thing. I have learned to keep a copy hanging around. (I am far from Word's limits: I only use aboput 6 special styles, plus redefining normal, the TOC styles, and the outline (level) styles. Why is the Microsoft hotline closed on weekend and holidays when everyone knows that is when all crashes occur? :-( Yes, once you get used to it, Word has a lot of things needed for a professional editor. If it would only be reliable. don norman Donald A. Norman Institute for Cognitive Science C-015 University of California, San Diego La Jolla, California 92093 norman@nprdc.arpa {decvax,ucbvax,ihnp4}!sdcsvax!ics!norman norman@sdics.ucsd.edu norman%sdics.ucsd.edu@RELAY.CS.NET
moku@sphinx.uchicago.edu (Mark Francillon) (05/26/87)
In article <365@sdics.ucsd.EDU> norman@sdics.UUCP (Donald A. Norman) writes: [document blows up] > >There is a Word Temp 44 file some 77K long. Word can't load it. > >In desperation, I change the TYPE and CREATOR on the Word Temp file to >be WDBN and MSWD and just try loading the temp file into Word 3. It >works! I only lose 2 hours of stuff (I back up every few minutes -- >why did I lose 2 hours?) > I've had to do this too. Nice that it works, but very risky business, since there's no telling -as far as I can see- when or whether Word's going to write out one of these temps (This also means the temp isn't just the last save, so saving every few minutes isn't necessarily going to help you here). I've found that the recovered file is sometimes still in a strange state (i.e., the original problem may reappear). Cutting the contents and pasting them into a new document seems to help. >The Table of Contents suddenly stopped working a week ago. That is, >it works fine on short files, bombs, consistently on long files. It >used to be an essential feature. Now it is useless. It seemed to Wait 'til you get to indexing. Special Memorial Day bomb-o'-the-day: open a file and split the screen (either a regular split screen or a footnote window will do); put the insertion point in the upper pane and click in the lower pane within an eighth of an inch or so of the bottom of the window. Cherry bomb =02 every time. Footnoters beware. Mark Francillon Dept. of Anthropology University of Chicago ...ihnp4!gargoyle!sphinx!moku
chuq%plaid@Sun.COM (Chuq Von Rospach) (05/26/87)
>>The Table of Contents suddenly stopped working a week ago. That is, >>it works fine on short files, bombs, consistently on long files. It >>used to be an essential feature. Now it is useless. It seemed to I suggest you try deleting your "word preferences" file and rebuilding it from scratch. There are a few bugs involved in how Word handles this file, and it clears up a large number of problems that carry over multiple executions of the binary or rebooting. >Special Memorial Day bomb-o'-the-day: open a file and split the screen >(either a regular split screen or a footnote window will do); put the >insertion point in the upper pane and click in the lower pane within >an eighth of an inch or so of the bottom of the window. Cherry bomb =02 >every time. Footnoters beware. I tried it on my machine. It doesn't fail. Again, try zapping "word Preferences" and see if that helps. If not, try it on a virgin file. If it still fails, there is something unique about your system (inits? da's? strange hacks?) that is interacting badly with what you're doing, because I don't have that problem... Which just goes to show -- before you go blaming a program, try to find out if it is really the program that does it. Try it on the distribution floppies, for instance -- with all the neat toys folks put in the System file/folder, there are lots of ways to screw up a system that you might not even be aware of. Just because Word has bugs doesn't mean that ll of the bugs in a Mac are in Word. Before you start passing blame, do your homework. chuq (official bug catcher type person) Chuq Von Rospach chuq@sun.COM [I don't read flames] There is no statute of limitations on stupidity
moku@sphinx.UUCP (05/26/87)
In article <19647@sun.uucp> chuq@sun.UUCP (Chuq Von Rospach) writes: > >>Special Memorial Day bomb-o'-the-day: open a file and split the screen >>(either a regular split screen or a footnote window will do); put the >>insertion point in the upper pane and click in the lower pane within >>an eighth of an inch or so of the bottom of the window. Cherry bomb =02 >>every time. Footnoters beware. > >I tried it on my machine. It doesn't fail. Again, try zapping "word > >Which just goes to show -- before you go blaming a program, try to find out >if it is really the program that does it. Try it on the distribution >floppies, for instance -- with all the neat toys folks put in the System >file/folder, there are lots of ways to screw up a system that you might not >even be aware of. Just because Word has bugs doesn't mean that ll of the >bugs in a Mac are in Word. Before you start passing blame, do your homework. > I agree in principle as to carelessly made charges. We all know that reputations are more easily damaged than repaired. I wouldn't have said anything if I hadn't been able to replicate the bomb on several different machines, differently configured --including time after time in front of a dismayed Microsoft rep at a MacFest here a couple of weeks ago. Try getting the tip of the arrowhead as close to the bottom of the lower pane as possible. Of course this makes the whole procedure sound contrived, but I assure you that if you spend any amount of time working with footnotes, this one will bite with regularity --unless you can drill it into your head not to try jumping to the bottom line of the footnote window. Mark Francillon Dept. of Anthropology University of Chicago ...ihnp4!gargoyle!sphinx!moku
norman@sdics.ucsd.EDU (Donald A. Norman) (05/26/87)
Here is what Microsoft said about my problems 1. The hangup in save. A known bug. Will be fixed in the June release. Yup, he said June. Use save-as: much safer. I asked him to repeat this. Save-as is safer. (hah. Only an hour after being told this, as I was dutifully saving-as, the system froze again. Fortunately, I had just a second earlier done a save-as to a floppy.) The APPLE file handler makes this awfully inconvenient APPLE, are you listening? If you want to save the same file on two different drives, you lose the directory path. When you return to the original drive, the file menu starts youoff at the root. Can't it retrieve the path last used for the device? 2. The problem with table of contents. Probably a defective finder. The finder can't access the disk properly. Throw away the finder and plug in a virgin one. I did as the man said. No luck whatsoever. 3. Recovering from Temp files. Sheer luck. Don't count on it. Temp files are written only at the time Word starts up. 4. Trashing the setup file. No hints. ----- A respondent suggested my problems were due to a perverse prefernce file and that I should rebuild it. The thought had already occured to me, but that didn't do the trick. -------- final score: Microsoft was very helpful and polite, but not a single one of their suggestions worked. don norman Donald A. Norman Institute for Cognitive Science C-015 University of California, San Diego La Jolla, California 92093 norman@nprdc.arpa {decvax,ucbvax,ihnp4}!sdcsvax!ics!norman norman@sdics.ucsd.edu norman%sdics.ucsd.edu@RELAY.CS.NET
anson@elrond.UUCP (05/26/87)
In article <365@sdics.ucsd.EDU> norman@sdics.UUCP (Donald A. Norman) writes: > A routine back-up save to the hard disk hangs up. 66% written >and nothing is happening. The 66% figure stays there, the clockface cursor >sits there, but no disk noises. j-clock keeps running. 20 minutes >later it is still there. So I take a few more deep breaths and reboot. > [interactions? j-clock. Superspool Imagewriter spooler. > Standard Mac +. DataFrame 20 Mbyte hard disk. Nothing else > that should cause problems.] This may or may not be the problem... I also use Superspool Imagewriter spooler with Word 3.0, and a curious interaction occurs. From time to time, I notice that Superspool has stopped printing out data, even though it is far from done. Simply clicking the mouse in a Word window starts it up again. This doesn't happen with other applications, and probably has something to do with other input related phenomena noted previously on the net. Could it be that Word and Superspool can deadlock during a save?(!!!) (The original poster didn't mention whether printout was in progress when this happened.) -- ===================================================================== Ed Anson, Calcomp Display Products Division, Hudson NH 03051 (603) 885-8712, UUCP: [decvax, wanginst, savax]!elrond!anson (Just blame me; my boss isn't even certain about just what I do.)
korn@cory.Berkeley.EDU.UUCP (05/27/87)
In article <365@sdics.ucsd.EDU> norman@sdics.UUCP (Donald A. Norman) writes: > >...[problems with MS-Word 3.0... speculates on possible problems...] > > [interactions? j-clock. Superspool Imagewriter spooler. > Standard Mac +. DataFrame 20 Mbyte hard disk. Nothing else > that should cause problems.] From what I've seen (several owners of Dataframes who use J-clock and SuperSpool) these two are a fairly deadly combination. Especially if you are using the older J-clock (a known Hard Drive killer). Get the new j-clock (or simply don't use it), and get SuperLaserSpool (which also spools to the imagwriter, is much cleaner, and is free to all DataFrame owners--bitch at your dealer if he doesn't give it to you free; he should). This may solve some of your problems... Peter ----- Peter "Arrgh" Korn Hacker? Me? A hacker? No, actually korn@ucbvax.Berkeley.EDU I'm a mac-er. All's we do is {decvax,dual,hplabs,sdcsvax,ulysses}!ucbvax!korn make library calls.
johnson@su-russell.UUCP (05/27/87)
.... I have experienced the ``Freeze'' bug that Don Norman mentions. It shows up when working on Word 1.05 files converted to 3.0 files. Mark Johnson Center for the Study of Language and Information
jww@sdcsvax.UCSD.EDU (Joel West) (05/28/87)
I've also seen long pauses at 66%. No crashes yet, but the pause is longer than all the time so far and I wonder if it will recover... -- Joel West {ucbvax,ihnp4}!sdcsvax!jww (ihnp4!gould9!joel if I ever fix news) jww@sdcsvax.ucsd.edu if you must
ephraim@wang.UUCP (05/28/87)
In article <367@sdics.ucsd.EDU>, norman@sdics.ucsd.EDU (Donald A. Norman) writes: > Here is what Microsoft said about my problems > ... > 2. The problem with table of contents. Probably a defective > finder. The finder can't access the disk properly. Throw away the > finder and plug in a virgin one. > > I did as the man said. No luck whatsoever. > No reason it should work, since the Finder isn't running when Word is. The Finder is "just another application" and has little to do with the behaviour of other applications. What your friendly, fuzzy, but slightly confused Microsoft help-line person probably meant to suggest was to replace the *System* file, which contains all the standard-file stuff and can have a profound effect on the apparent health of other software. Confusion between the Finder and the System is common among Mac users, but someone paid to help users should know better. Much better.
russ@oakhill.UUCP (Russell Schwausch) (05/29/87)
In article <365@sdics.ucsd.EDU> norman@sdics.UUCP (Donald A. Norman) writes: > A routine back-up save to the hard disk hangs up. 66% written >and nothing is happening. The 66% figure stays there, the clockface cursor >sits there, but no disk noises. j-clock keeps running. 20 minutes >later it is still there. So I take a few more deep breaths and reboot. > [interactions? j-clock. Superspool Imagewriter spooler. > Standard Mac +. DataFrame 20 Mbyte hard disk. Nothing else > that should cause problems.] I have been working on a 40 page manual that has grown to > 160K. I also noticed the pregnant pause at the 66% mark but it is usually just 2 or 3 minutes and then it goes on to complete the save. I thought it was because I had 4 pages full of graphics toward the end of the document but your experience hints that there is no relationship there. I also use jclock. However, I have seen the pause on both systems with hard disk and systems without. And I have never had a hang at the 66% mark. >There is a Word Temp 44 file some 77K long. Word can't load it. I have had Word 3.0 leave temp files of various sizes laying around many times. In fact, on my home system they seem to accumulate in the system folder (Word 3.0 is one level up) and I have to manually clear them out. A call to Microsoft got me the answer that "yep, it does this! Is a fix coming? dunno". This does not happen on every system I work on!!!???@#$%....... >In desperation, I change the TYPE and CREATOR on the Word Temp file to >be WDBN and MSWD and just try loading the temp file into Word 3. It >works! I only lose 2 hours of stuff (I back up every few minutes -- >why did I lose 2 hours?) Thanks for this info. It might come in real handy someday. >---- >QUESTIONS and OTHER PROBLEMS > >What is the purpose of the temp file? Why is it a backup? If it >really is a backup, why doesn't MS word provide a recovery process for >it. My guess is Word 3.0 is being very careful to save old file before replacing it with new file in case of a crash (I have had several after the 100% mark) during the save. My problems doing a save usually crop up when I don't have enough room on my disk(s) for 2 copies of the file I am editing. Then I try a Save As, pop a disk out, put a fresh disk in, and watch the system bomb instead of saving all my valuable work. >The Table of Contents suddenly stopped working a week ago. That is, >it works fine on short files, bombs, consistently on long files. I haven't had any problems with the toc and I regenerate it frequently on this 160K+ document I'm writing. I have noticed that it gets upset if I'm not careful when changing the from and to levels (I use 1 thru 7 for headings, 8 for illustrations and 9 for tables; works fine). I use .c1, etc. rather than outline to designate toc entries. >And every so often Word reverts to innocence. I boot the system and >Word displays short menus. No specially tailored menus. No memory of >all my styles. The Word Settings File is clobbered. Why? No clues. This happened to me early on when Word got confused about where to keep the settings file. Sometimes is was at the same level as Word 3.0 application and sometimes it was in the system folder!!! My system was also having trouble finding the application unless I had the volume it was in open on the desktop. I rebuilt the desktop (hold down cmd & option keys while quiting an application) and had no more problems with finding the application or the settings file. >Yes, once you get used to it, Word has a lot of things needed for a >professional editor. If it would only be reliable. Amen to both. I'm keeping my fingers crossed until they cut loose the update that is supposed to be coming the end of June. > >don norman -- Russell Schwausch, Motorola Inc., OakHill, Tx. (A suburb of Austin) UUCP: {harvard,ihnp4,seismo,gatech,nbires}..!ut-sally!oakhill!russ Ma Bell: (512)440-2426 It's a ticket agent, it's a fare collector, no ... it's SUPERCONDUCTOR!!!
russ@oakhill.UUCP (Russell Schwausch) (05/29/87)
In article <1825@sphinx.uchicago.edu> moku@sphinx.UUCP (Mark Francillon) writes: >In article <365@sdics.ucsd.EDU> norman@sdics.UUCP (Donald A. Norman) writes: > >>The Table of Contents suddenly stopped working a week ago. That is, >>it works fine on short files, bombs, consistently on long files. It >>used to be an essential feature. Now it is useless. It seemed to > >Wait 'til you get to indexing. I have not had problems with either table of contents or indexes (indices, whatever) on a long document > 160K. I suspect you have inadvertently fouled up a ".cnHeading;" and/or ".i.word;" entry in your document and are trying to collect a verrrrry BIG heading or index entry. Of course I have accidently deleted a semicolon on an index entry and gotten the rest of the paragraph in my index with no bomb. But when I don't mess up, things work fine. I have about 45 entries in my toc and 120 entries in the index. -- Russell Schwausch, Motorola Inc., OakHill, Tx. (A suburb of Austin) UUCP: {harvard,ihnp4,seismo,gatech,nbires}..!ut-sally!oakhill!russ Ma Bell: (512)440-2426 It's a ticket agent, it's a fare collector, no ... it's SUPERCONDUCTOR!!!
mayerk@linc.cis.upenn.edu (Kenneth Mayer) (05/29/87)
I have had a document die on me too. My solution was to change it to a TEXT file. Of course, all formatting goes out the window, but the text was mostly recovered. Kenneth Mayer: mayerk@eniac.seas.upenn.edu
fisher@elxsi.UUCP (Chuck Fisher) (06/01/87)
I too had a bad experience with Word 3.0 over the weekend. I had spent two hours taking notes at a meeting with the Mac and had about 6 pages of text. All the while I was saving frequently (about every 15 minutes) and never a hint of trouble. Untile I returned home, that is. I tried to open the document and I get the infamous bomb ID 02 (bad address). After several unsuccessful ruses (turning off cache, switching documents, copying the bad document, etc.) I decided to try one more trick. I made a duplicate of the bad document and changed its type to "TEXT" with an originator of "????". I figured that should fool Word 3.0, and it did! I opened the "text" document and everything was fine, except for no format information. It was an easy process to cut and paste to a new document and apply the styles that I had defined. It only took about 10 minutes to resurrect the thing. So far I have not had a repeat of the problem. I hope this helps some other people out in net-land. Chuck
mayerk@linc.cis.upenn.edu (Kenneth Mayer) (06/02/87)
Considering all of these horror stories about crashed documents, I wonder why Microsoft didn't think to build in a recovery routine in 3.0. That would make it a sensible design that crashes gracefully with the minimum amount of damage. Considering the reports, it seems to be a simple routine: Copy the file. Try to open it as a formatted document, and if that doesn't work, open it is a textfile. (But who ever said MS had *sensible* designs :-) Kenneth Mayer: mayerk@eniac.seas.upenn.edu
briand@tekig4.UUCP (06/02/87)
With regard to the Word 3.0 file destruction during save, be aware that ususally the program will save only the changes to a file in an incremental mode. At some level of complexity, Word 3.0 will decide on its own to do a complete save. I have never (yet!) lost a file, and I believe it is because I always force a full save. This can be done by doing a Save As, and noting whether the Fast Save checkbox control is checked. If it is dimmed, Word 3.0 will only do a full save, which is what you want. If Fast Save is live and checked, Word 3.0 is about to do an incremental save. I know what you're thinking: "On a Save AS?!?" Anyway, uncheck the Fast Save box and save the file under its old (default) name. You will be asked to verify the overwrite. There is some pain associated with this, and of course the full save is somewhat slower, but if it really makes saving more reliable it is worth it until MS gets its act together. I would be very interested in hearing from people to see if this prevents the problem. -- -Brian Diehm (SDA - Standard Disclaimers Apply) Tektronix, Inc. briand@tekig4.TEK.COM or {decvax,cae780,uw-beaver}!tektronix!tekig4!briand