Actually it makes sense, in a Microsoft sort of way. In the effort to make its WYSIWYG editors as WYSIWYG as possible, it offloads some rendering to printer drivers so as to mimic the printed copy as closely as possible. XPS has no rendering standard so it uses sane (but not good) defaults.
Of course they've never heard of PostScript. This kind of brain-damage is just a tiny part of the failure that is Microsoft WYSIWYG and typesetting "technology".
With a proper typesetter like LaTeX you get a PDF that's a dot-for-dot match with what you'd get in a calibrated printer, without ever having to assume any particular printer. It's the printer's responsibility to implement PostScript properly, not the typesetter's responsibility to tune its PostScript to the printer!
OMG. This problem is old... I do not know how old it. It was so many years ago.
With some NICs under Win9x one had to do some hand waving to make it working. And two reboots. (Good NICs with good OEM drivers (e.g. Intel) had no the problem - setup.exe did it all for you. But e.g. RealTek was shipping only drivers, without any fancy installation program.) I already forgot what to do precisely, but yes, it was caused by Win9x not installing something during setup since network wasn't present (but some dummy stuff was installed instead). The installation of missing pieces could be triggered artificially later - with minimum two reboots - but how and what were the step I already forgot. Haven't seen Win9x for 10+ years now....
I still remember though the impression of people when I did extra redundant reboot and Win9x network was magically coming to life. (*)
(*) Not always, as Win9x's DHCP/WINS was atrocious and sometimes also causing the effect as if network was down.
but it was back in the days of Windows 95. I was working in software Localisation for a Lotus Notes product. We had several machines working in the test lab based on ghost images, so they were all pretty much identical.
One of the machines kept dying on us during the test phase, but none of the others did. Very confusing, for about a day. Until we realised that the machine which was crashing had an audio CD in the drive. (Not playing, not in Explorer. Just present in the drive.)
We verified it by swapping the audio cd into other machines, and running the same tests. Invariably, the machine with the CD in, crashed when we tried to perform task "x" in Lotus Notes.
It was escalated up, as I recall. And we eventually got a note back saying "Don't put CD's in the CD-Rom drives."
I still remember it (as a recent graduate) as my first exposure to management-style thinking.
Depending on what else they did, that might be a good response. A proper IT service desk should do two things in a situation like this:
1. It should find a quick workaround for the incident at hand, which is to recomend all customers to not put an audio CD in the drive of a server running notes.
2. The should perform root cause analysis to determine the underlying problem and remove it permanently.
If the Service desk isn't doing both these things, it's not doing its job properly.
I complained to Dell recently about the optical drive on my new laptop. If you put in a DVD with the slightest scratch on it the whole system would hard lock. (Yes, Vista allows the drivers the privilege of taking down the whole OS.) Anyway, the support dude was like "well, duh, we can't read a scratched disk." More googling revealed that TCorp had released firmware to fix the problem.
I recently filed a bug about a certain popular grid control's mouse wheel behavior. The company making the control responded that it was not a bug because "Microsoft's [ancient] grid control has the same behavior." Gee thanks, dorks. Good thing you set your standards so high.
Double click on a document. Word sits there for what seems like hours saying something like "Connecting to default printer. Press ESC to stop" so you give up and press ESC and start editing the document. Word promptly crashes. The workaround - set the default printer to Microsoft XPS and select the printer manually when you need it and wait the eternity it takes to communicate with the network printer. And sometimes it crashes again. WTF?
WinXP has issues connecting to Win98 SMB printers via TCP or NetBEUI when connected to a DOS6 network running LANtastic. It would take about 15 minutes to find the printer and about 10 minutes to send a small document. There was no problem browsing the network, though.
LANtastic had some suggested workarounds (changes to how broadcast packets are routed by LANtastic nodes and changes to the TCP and SMB configs in Win98, mostly involving registry hacks), but it turns out the only reliable workaround I found was to install an lpd emulator on Win98 that connected locally to the printer, and then have WinXP connected to lpd. It worked quite reliably and was quicker at connecting than I'd ever seen an SMB printer be. That wasn't an official workaround, though, just something I tried on a hunch.
I remember in the early days of libtool... depending on what version of automake tools were included in a package, what version of the automake tools you had elsewhere on your system, your version of libc, the version of bash you used, the versions of make and gcc you had installed, and the veerssion of text-utils and sh-utils you had, sometimes libtool would generate very long command strings with hundreds of redundant arguments, and then call itself to "simplify" the arguments but actually recurse with an even longer string, until bash segfaulted and your login session crashed.
There was never really a workaround for ttha... just "try different veersison of thinggs, you might needto downgrade automake, or mix and match different veersison of auttoocnf, automake, and libtool." Quite wonderful, I tell you.
gcc2.7.2.3 (the really stable version you had to compile the linux kernel with for quite some time) had some weird bug that didn't really have an official workaround, either. Somehow if you did pointer calculations on the function argument list (like varargs or stdarg) andn the called another function, the last local variable of the called function couldn't be written until it was read. I remember having to do something like printf("", a); before a statement like a=4; would work. Of course, then you'd get a warning about using an uninitialized variable, but... The funny thing was, I seem to recall that only would happen when optimizations were turned *off*. Turning them on made the bug go away, which made it really frustrating to track down. It ended up being something like gcc subtracting the wrong multiple of 4 from the stack pointer (under all the aforementioned conditions) in the block of asm that set up the stack frame. Of course, gcc2.81 and 2.95.2 had their own issues, and egcs wasn't much better... It wasn't until gcc3.2 where I didn't need multiple versions of gcc (one for the kernel, one for the program I was working on, and one that compiled c++ code correctly!!)
I remember MatlabR11 having broken CSV-file-parsing routines. The workaround? Write your own. The Matlab compiler was also moving to a new system (MEX), but there were lots of things that didn't work yet, and the previous compiler system was officially deprecated. Then, the next release of Matlab required 92MiB of DLLs to be installed as a Matlab runtime if you wanted to distribute anything you compiled with the Matlab compiler... and much of that runtime was broken Java libraries. A lot of the official suggestions for working with structured data that involved strings required many layers of nested cell objects, which had their own compilation issues. Again, the workaround was to convert string tables into padded numeric matrices of UInts. Of course, most of the matrix manipulation functions only worked with Real numbers, so you had to convert back and forth, and be careful about what type of rounding/flooring/ceilinging you were doing...
VB6 had a broken val() that returned the wrong values for ASCII characters in the range 160 through 184 (I think),, butthere wasn't realalyy n conssitent pattern. MSDN and the Microsoft KB gavee th official workaround: write your own val().
I had an older version of Word and I wanted to make an A3 document - but my printer only supported A4.
You're lucky in that you appear to live in a locale that uses ISO 216 (A-series) paper sizes. ISO paper, unlike the U.S. letter series, has a nice mathematical definition: all sizes are the same aspect ratio of sqrt(2):1, and each size has twice the area and sqrt(2) times the length and width of the size below it. So make your document on A4 and print it on A3 at 141%.
In March, the Google Docs team introduced the Drawings feature. Now you can create drawings, schematics etc. in your Google Docs document. Now when you want to print your doc, or export it to some other format than HTML, then you get a nice error message [google.com].
If you want to export or print, the workaround for the last three months has been... not to use drawings in your documents! Great feature!
20 years ago I worked with an application on VMS. It used some form based UI tool which you get with the OS. (was it ACMS? I can't remember now) anyway you could set a timeout on a form which kicked you back to another screen if you didn't complete it within a specified time. One form with 20 fields or something had a timeout of ten seconds. There was something strange about the guy who wrote that...
Specifically, go down to near the bottom of the entry where it mentions that: [t]he equipment control task did not properly synchronize with the operator interface task, so that race conditions occurred if the operator changed the setup too quickly. This was missed during testing, since it took some practice before operators were able to work quickly enough for the problem to occur.
The Therac-25 incident also includes a great example of a ridiculous workaround for a serious (fatal!) software bug, the race condition triggered by this fast typing, or using an unexpected sequence of keys. The manufacturer's initial suggested fix was:
"Effective immediately, and until further notice, the key used for moving the cursor back through the prescription sequence (i.e., cursor "UP" inscribed with an upward pointing arrow) must not be used for editing or any other purpose.
To avoid accidental use of this key, the key cap must be removed and the switch contacts fixed in the open position with electrical tape or other insulating material. For assistance with the latter you should contact your local AECL service representative."
Quite rightly, the FDA concluded this was completely inadequate:
Lotus Domino server installations (circa 2000) would complete at about four to five times their "normal" speed if someone just sat there moving the mouse around whilst the install wizard was copying files. Go figure.
I'd suggest trying the hates-software website at we.hates-software.com, but the software crapped out over a year ago and the guy running the site can't be arsed tracking down the no doubt obscure bug in Mariachi and fixing it. Since all of the users are too busy hating software they have to work with to fix software they're not actually responsible for, it's probably never going to get fixed, which is hateful but somehow satisfying, in a kind of Zen way.
I quite like the workaround that's always given for content management systems that can't strip out the humongous amount of invisible HTML cruft that comes with text that's copied to the clipboard from MS Word or Outlook.
Content editor: "Hey, why is the formatting of this page completely borked? And why can't I use the CMS's editor to fix the borkage?"
Me: "Where did you get the original text from?"
Content editor: "I copied it from a Word doc that somebody sent me. I just pasted that in. It was just plain text..."
Me: "I see. Well, delete the page and start again. This time, copy the stuff from Word, then open Notepad, past the text from Word into Notepad, then copy/paste into the CMS from there instead."
I remember when Microsoft put a crappy 32-bit front-end on MS-DOS 7.0 to make it more useful. It completely sucked. It hogged memory and crashed all the time. Luckily you could boot directly into DOS to avoid the GUI and get real work done.
Many years back I was working as a freelancer developing the training material for a customer service app.
The agents input customer details, the app identified the nearest call-out contractor, sent the contractor a text message, started the clock ticking and updated the log.
Unfortunately, the devs used their own GUI and in the top row the 'submit' button was right next to 'form clear' and call centre staff kept clicking the wrong button, erasing the customer details and having to ask for them all again. This did not go down well with customers who'd called due to a domestic emergency (plumbing etc.)
I suggested that the workflow through the form meant that the agents would be better served by a submit button at the bottom.
The response to my submission: "Can't see a need to move the button during this development cycle - agents to be told to stop clicking the wrong button."
seems like an obvious feature it should have shipped with. A product called Offline Review for a medical imaging device for a cancer treatment system. The problem: it shipped before the "offline" part was implemented. Recommended workaround: have the physician available to review the image during the treatment rather than on his own time. Yeah, because physicians can stop having clinical hours so that they can watch each treatment that therapists' do, and oh yeah patients from the same doc have to be secheduled at different times to allow for this. Nice.
When you used a computer as a time clock (running the client software and using a card swiper, instead of buying the special timeclock hardware), the licensing system on the "server" (which had to be logged in to run, as it wasn't a service but a running process) would lose track of a particular computer's license if more than one computer was running the timeclock client - and issue a new one the next time the client was run.
So, if you had purchased 15 licenses and were running 2 or more clocks (but less than the 15 you were supposedly allowed), you'd run out of licenses after a couple of days, even with light use.
After working for a month or so with the company to resolve the issue, what was their long term solution?
Give us a code that would give us "unlimited" (or somewhere in the area of 32,000 licenses).
After several years (like 8 or so) and much griping from me to switch to something else, we're still using the software, actually (but with only one swipe station, and only for our student workers in our biggest department), but will supposedly switch to something hosted and web based "soon".
I guess many would be aware of the case of the 500-mile email. An office was not able to send emails to places which were physically located at a distance greater than 500 miles from the office! Entire story and the logic behind it can be read here - http://www.ibiblio.org/harris/500milemail.html [ibiblio.org]
How about Ubisoft and RB6 Vegas? Remember that their fix around a big DRM issue was basically to install a no-cd crack by Reloaded? They just took the crack, renamed it, and then released as an official patch.
These weren't official ones. I developed them on my own. The tape recorder was notoriously difficult to get the data to load right. Some tapes, saved on a different recorder, would require special tricks to get the readout "within specs".
One, I had to mute audio in the TV set to which the Atari was hooked up. I guess electromagnetic interference from the speaker was a problem. On another, I'd have to hold the label with key functions on the recorder. The label was metal and connected to the recorder ground. By holding it, I was providing extra grounding that reduced the noise just enough to get the game to load. Luckily that one took only like 5 minutes to load:) The best one was copied from a floppy. The copy was good, but there was no 'loader' program and the game was too big to fit with a copier to copy it to a different tape, and recorded from the beginning of the tape, no room to save the loader. The solution was to take a random different tape with a generic loader, start loading it, then after counting 6 "beeps" QUICKLY remove it and put the right tape in - the timeout tolerance was like 2-3s, so you really had to hurry.
Way back when dinosaurs roamed the earth, ok not really, but in the late 80s, Sun had a problem with some of their hard drives. When they would park they heads they would stick and you couldn't unpark them. Sun's solution was to tell you to HIT the computer. They even sent us a letter showing you where on the "pizza box" enclosure one would strike.
After upgrading a server, we watched a client verify the server through his daily application. The client entered data and clicked on submit, the next screen appeared instantly. "This is not possible" said the client "it takes about two seconds to submit data to the database"!
"But the new server is much faster!" we said. It didn't matter, the client refused to believe the data was really submitted.
We held a meeting about this 'problem'. One developer suggested to add a two second 'do nothing' loop to the submit button.
So we patched the server and asked the client to verify again. He entered data, clicked 'submit' and was very happy to have his two second delay back! "Now it works..." he said "...now the data is entering the database!".
We admitted our fault (knowing very well that all we added was a two second delay).
I used to manage Digital UNIX (later called Tru64) systems for a large, now bankrupt, telecom back around the turn of the millennium. The filesystem used, AdvFS, was pretty cool and advanced for the time but under the version of the OS we were running we found that free space would shrink at a faster rate than used space would grow. I had filesystems report full even though a df would show only 60% used.
It turned out that when small files were deleted all of the space wouldn't become free. My customer wrote thousands upon thousands of 150-200 byte files a day and deleted just as many. The entire team and my customer agreed this was clearly a bug.
When brought up with Compaq (who had recently aquired Digital) the technical rep investigated and reported "this is not a bug, the code is being executed exactly how it's written." Seriously, this was his response. I would have been more amused if he seriously argued it was a "feature."
I never could get a definition of what a "bug" really was from him. I became rather infuriated when he reported to me that this issue was "fixed" in the latest major release of the OS. If there was no bug, why was it fixed?
I never got a straight answer and was left on my own to find my own work-around which involved inserting a new volume into the filesystem thus growing it and then deleting an old volume. When this was done to all volumes in the filesystem, the problem was resolved for a few more months. This was an incredibly labor intensive and, as far as I'm concerned, incredibly risky to move data around like that on a hot system with insane uptime requirements. There was also a massive performance hit while this was happening and my customer's application was already VERY IO intensive.
I'm still just as angry about that conversation with the rep today as I was back then.
I will say that self documenting words (just like self documenting code) require a minimum intelligence level. I'm wondering what percentile of the US population you represented to get the "fan fucking" + "elastic" conclusion.
I don't understand how people associate the word "fuck" so exclusively with sexual meaning. It seems to be a more prevalent attitude in America, affecting even supreme court justices.
"Even when used as an expletive, the F-word's power to insult and offend derives from its sexual meaning," Scalia said.
Such a conclusion is a pretty unfair typecasting of such a versatile swearword. While "fucking" or "to fuck" is often used to describe sexual intercourse, the word has a great many other meanings. "Fuck off" being the most classic and familiar example, used to gruffly tell someone to remove themselves, or to desist from an action, etc, but perhaps only to express disbelief or some such. "What the fuck" shows the ability to use the word in an undirected fashion. Alone, "Fuck" can be an effective emotional outlet. "Fuckers" turns the verb into a noun, that is, if it were ever a verb in the first place. Things like "fan-fucking-tastic" show just how versatile this unique utterance can be, as it transcends classical descriptions.
So, "Fuck" is not just a sexual swearword. Perhaps, lacking any other terms, American's take it to primarily refer to intercourse. In fact, other english speakers have many other words at their disposal for describing sexual activities. "Shag","ride", etc. Lack of such words in someones personal or cultural lexicon should not be used to imbue unwarranted meaning to a speakers words in some kind of reverse irony.
When Bono said "fucking brilliant" at the Golden Globes, it was clear to any reasonable person that he meant the word as an adjective to brilliant, not as a sexual reference. This is doubly clear to anyone from Ireland. Nevertheless the FCC claimed that the word had and "inherently has a sexual connotation", in any context. And worse, the US supreme court agreed with them.
As someone who has been told on countless occasions by friends, family and countrymen to "Fuck off", or some such like, I'm personally offended far more by the suggestion that all these people's comments had an underlying sexual meaning than I am by any of the expletives themselves. But once again I find my culture, my traditions, my airwaves, and my internets subjected to the interpretations and censorship of conservative bible bashers in rural America. It's fairly insulting.
So please accept my sincerity when I say that you, and all those that would corral honest swearwords into narrow definitions, respectfully, Can All Fuck off with Yourselves!
Do you have any idea how hard random data is to collect?
If your PC has a sound card, an entropy gathering service can hash the microphone input and derive at least 1 high-quality random bit per sample from ADC dither noise alone. So that's 96 kbps for a typical 48 kHz stereo ADC.
I think this actually had a good reason. A nice old PS2 mouse generates interrupts when wiggled. This breaks up the boring routines. (Blocking routines actually.) And presto, a little more progress on transfering your data...
This phenomenon is not gone btw. 1. Start notepad in a window, not full screen. 2. Open long text file 3. Mark your text from beginning of document and try to scroll down. When mouse exits window, keep holding but with mouse stationary. Nothing happens? 4. Wiggle mouse outside window and presto it continoues to mark text towards the bottom of your document!!!
John Wayne's not dead - he's frozen! And when we find a cure for cancer, we're gonna thaw out the Duke and he's gonna be pretty pissed off. You know why? You ever taken a cold shower? Well, multiply that by 15 million times. That's how pissed off the Duke's gonna be.
speaking of HP printers, especially the networked ones, why is it that the network driver is 350 megs in size? I had to download two of those damn things, even after using a custom install option, to remove as much of the cruft as possible I still installed some 700 megs of drivers for two printers, and a scanner.
Guess what happens when the drivers get corrupted. you have to manually uninstall the registry settings and deleted all files manually in order to reinstall the drivers or they won't work.
HP decent printers, Software coded by monkey banging on keyboards.
Run Windoze much?? (Score:5, Insightful)
HP and Microsoft repeatedly suggest re-installing the operating system to cure a network configuration issue.
Re:Run Linux much? (Score:5, Informative)
Funny, I've had people tell me to reinstall the new Linux(here, uBuntu) updated set instead of updating it.
Maybe I'm a bad luck magnet, but last time I tried to update it pulverized X.
With apologies to Staples:
"That Was Fun!"
Parent
Re:Run Linux much? (Score:5, Insightful)
Oh, 9.04 was crap and everybody knows it. At least on the Intel driver front, and that's just for starters.
They said that about 8.10, and 8.04.
Parent
Re:Don't plug in your scanner! (Score:5, Interesting)
Actually it makes sense, in a Microsoft sort of way. In the effort to make its WYSIWYG editors as WYSIWYG as possible, it offloads some rendering to printer drivers so as to mimic the printed copy as closely as possible. XPS has no rendering standard so it uses sane (but not good) defaults.
Of course they've never heard of PostScript. This kind of brain-damage is just a tiny part of the failure that is Microsoft WYSIWYG and typesetting "technology".
With a proper typesetter like LaTeX you get a PDF that's a dot-for-dot match with what you'd get in a calibrated printer, without ever having to assume any particular printer. It's the printer's responsibility to implement PostScript properly, not the typesetter's responsibility to tune its PostScript to the printer!
Parent
Re:Run Windoze much?? (Score:5, Informative)
OMG. This problem is old ... I do not know how old it. It was so many years ago.
With some NICs under Win9x one had to do some hand waving to make it working. And two reboots. (Good NICs with good OEM drivers (e.g. Intel) had no the problem - setup.exe did it all for you. But e.g. RealTek was shipping only drivers, without any fancy installation program.) I already forgot what to do precisely, but yes, it was caused by Win9x not installing something during setup since network wasn't present (but some dummy stuff was installed instead). The installation of missing pieces could be triggered artificially later - with minimum two reboots - but how and what were the step I already forgot. Haven't seen Win9x for 10+ years now....
I still remember though the impression of people when I did extra redundant reboot and Win9x network was magically coming to life. (*) (*) Not always, as Win9x's DHCP/WINS was atrocious and sometimes also causing the effect as if network was down.
Parent
RE (Score:5, Funny)
Re:RE (Score:5, Funny)
There, fixed that for you.
Parent
Re:RE (Score:5, Insightful)
There should be a +1, Sad But True.
Parent
Profiling? (Score:5, Informative)
A profiler was crashing when I tried to find bottlenecks in my code. The support rep. told me I should turn off optimization.
Profiling /= Debugging (Score:5, Insightful)
Profiling has to be done with same flags enabled as for the production code. Otherwise the result will be meaningless.
Parent
Re:Profiling /= Debugging (Score:5, Funny)
"Then don't optimize your production code."
Ticket closed.
Parent
Don't have the details (Score:5, Interesting)
but it was back in the days of Windows 95. I was working in software Localisation for a Lotus Notes product. We had several machines working in the test lab based on ghost images, so they were all pretty much identical.
One of the machines kept dying on us during the test phase, but none of the others did. Very confusing, for about a day. Until we realised that the machine which was crashing had an audio CD in the drive. (Not playing, not in Explorer. Just present in the drive.)
We verified it by swapping the audio cd into other machines, and running the same tests. Invariably, the machine with the CD in, crashed when we tried to perform task "x" in Lotus Notes.
It was escalated up, as I recall. And we eventually got a note back saying "Don't put CD's in the CD-Rom drives."
I still remember it (as a recent graduate) as my first exposure to management-style thinking.
Re:Don't have the details (Score:5, Insightful)
Parent
Re:Don't have the details (Score:5, Informative)
Depending on what else they did, that might be a good response. A proper IT service desk should do two things in a situation like this:
1. It should find a quick workaround for the incident at hand, which is to recomend all customers to not put an audio CD in the drive of a server running notes.
2. The should perform root cause analysis to determine the underlying problem and remove it permanently.
If the Service desk isn't doing both these things, it's not doing its job properly.
Parent
Re:Don't have the details (Score:5, Insightful)
Parent
Re:Don't have the details (Score:5, Insightful)
Because you can't sell bug fixes, only new features!!!
Parent
Re:Don't have the details (Score:5, Interesting)
I recently filed a bug about a certain popular grid control's mouse wheel behavior. The company making the control responded that it was not a bug because "Microsoft's [ancient] grid control has the same behavior." Gee thanks, dorks. Good thing you set your standards so high.
Parent
Stupid MS Office 2007 bug (Score:5, Interesting)
Double click on a document. Word sits there for what seems like hours saying something like "Connecting to default printer. Press ESC to stop" so you give up and press ESC and start editing the document. Word promptly crashes. The workaround - set the default printer to Microsoft XPS and select the printer manually when you need it and wait the eternity it takes to communicate with the network printer. And sometimes it crashes again. WTF?
Re:Stupid MS Office 2007 bug (Score:5, Interesting)
WinXP has issues connecting to Win98 SMB printers via TCP or NetBEUI when connected to a DOS6 network running LANtastic. It would take about 15 minutes to find the printer and about 10 minutes to send a small document. There was no problem browsing the network, though.
LANtastic had some suggested workarounds (changes to how broadcast packets are routed by LANtastic nodes and changes to the TCP and SMB configs in Win98, mostly involving registry hacks), but it turns out the only reliable workaround I found was to install an lpd emulator on Win98 that connected locally to the printer, and then have WinXP connected to lpd. It worked quite reliably and was quicker at connecting than I'd ever seen an SMB printer be. That wasn't an official workaround, though, just something I tried on a hunch.
I remember in the early days of libtool... depending on what version of automake tools were included in a package, what version of the automake tools you had elsewhere on your system, your version of libc, the version of bash you used, the versions of make and gcc you had installed, and the veerssion of text-utils and sh-utils you had, sometimes libtool would generate very long command strings with hundreds of redundant arguments, and then call itself to "simplify" the arguments but actually recurse with an even longer string, until bash segfaulted and your login session crashed.
There was never really a workaround for ttha... just "try different veersison of thinggs, you might needto downgrade automake, or mix and match different veersison of auttoocnf, automake, and libtool." Quite wonderful, I tell you.
gcc2.7.2.3 (the really stable version you had to compile the linux kernel with for quite some time) had some weird bug that didn't really have an official workaround, either. Somehow if you did pointer calculations on the function argument list (like varargs or stdarg) andn the called another function, the last local variable of the called function couldn't be written until it was read. I remember having to do something like printf("", a); before a statement like a=4; would work. Of course, then you'd get a warning about using an uninitialized variable, but... The funny thing was, I seem to recall that only would happen when optimizations were turned *off*. Turning them on made the bug go away, which made it really frustrating to track down. It ended up being something like gcc subtracting the wrong multiple of 4 from the stack pointer (under all the aforementioned conditions) in the block of asm that set up the stack frame. Of course, gcc2.81 and 2.95.2 had their own issues, and egcs wasn't much better... It wasn't until gcc3.2 where I didn't need multiple versions of gcc (one for the kernel, one for the program I was working on, and one that compiled c++ code correctly!!)
I remember MatlabR11 having broken CSV-file-parsing routines. The workaround? Write your own. The Matlab compiler was also moving to a new system (MEX), but there were lots of things that didn't work yet, and the previous compiler system was officially deprecated. Then, the next release of Matlab required 92MiB of DLLs to be installed as a Matlab runtime if you wanted to distribute anything you compiled with the Matlab compiler... and much of that runtime was broken Java libraries. A lot of the official suggestions for working with structured data that involved strings required many layers of nested cell objects, which had their own compilation issues. Again, the workaround was to convert string tables into padded numeric matrices of UInts. Of course, most of the matrix manipulation functions only worked with Real numbers, so you had to convert back and forth, and be careful about what type of rounding/flooring/ceilinging you were doing...
VB6 had a broken val() that returned the wrong values for ASCII characters in the range 160 through 184 (I think),, butthere wasn't realalyy n conssitent pattern. MSDN and the Microsoft KB gavee th official workaround: write your own val().
Early versions of t
Parent
You're lucky you use sqrt(2) sized paper (Score:5, Interesting)
I had an older version of Word and I wanted to make an A3 document - but my printer only supported A4.
You're lucky in that you appear to live in a locale that uses ISO 216 (A-series) paper sizes. ISO paper, unlike the U.S. letter series, has a nice mathematical definition: all sizes are the same aspect ratio of sqrt(2):1, and each size has twice the area and sqrt(2) times the length and width of the size below it. So make your document on A4 and print it on A3 at 141%.
Parent
Google Docs (Score:5, Interesting)
In March, the Google Docs team introduced the Drawings feature. Now you can create drawings, schematics etc. in your Google Docs document. Now when you want to print your doc, or export it to some other format than HTML, then you get a nice error message [google.com].
If you want to export or print, the workaround for the last three months has been... not to use drawings in your documents! Great feature!
Veterinary Clinic App (Score:5, Funny)
Oh yes:
We run a database-oriented app in a number of branches. It's so flaky that runtime errors are a daily occurrence.
The devs' response to reports of errors is usually:
a) Defrag the disk.
b) Stop the users typing so fast.
Seriously!
Re:Veterinary Clinic App (Score:5, Funny)
Parent
Re:Veterinary Clinic App (Score:5, Interesting)
b) Stop the users typing so fast.
Typing too fast caused people to die, in one case:
http://en.wikipedia.org/wiki/Therac-25 [wikipedia.org]
Specifically, go down to near the bottom of the entry where it mentions that: [t]he equipment control task did not properly synchronize with the operator interface task, so that race conditions occurred if the operator changed the setup too quickly. This was missed during testing, since it took some practice before operators were able to work quickly enough for the problem to occur.
Parent
Re:Veterinary Clinic App (Score:5, Interesting)
The Therac-25 incident also includes a great example of a ridiculous workaround for a serious (fatal!) software bug, the race condition triggered by this fast typing, or using an unexpected sequence of keys. The manufacturer's initial suggested fix was:
"Effective immediately, and until further notice, the key used for moving the cursor back through the prescription sequence (i.e., cursor "UP" inscribed with an upward pointing arrow) must not be used for editing or any other purpose.
To avoid accidental use of this key, the key cap must be removed and the switch contacts fixed in the open position with electrical tape or other insulating material. For assistance with the latter you should contact your local AECL service representative."
Quite rightly, the FDA concluded this was completely inadequate:
http://courses.cs.vt.edu/~cs3604/lib/Therac_25/Therac_3.html [vt.edu]
Start here for the whole sorry story:
http://courses.cs.vt.edu/~cs3604/lib/Therac_25/Therac_1.html [vt.edu]
Parent
Mouse wiggling not that unusual, surprisingly (Score:5, Interesting)
It's almost Zen. (Score:5, Funny)
I'd suggest trying the hates-software website at we.hates-software.com, but the software crapped out over a year ago and the guy running the site can't be arsed tracking down the no doubt obscure bug in Mariachi and fixing it. Since all of the users are too busy hating software they have to work with to fix software they're not actually responsible for, it's probably never going to get fixed, which is hateful but somehow satisfying, in a kind of Zen way.
Two-stage Pasting (Score:5, Funny)
I quite like the workaround that's always given for content management systems that can't strip out the humongous amount of invisible HTML cruft that comes with text that's copied to the clipboard from MS Word or Outlook.
Content editor: "Hey, why is the formatting of this page completely borked? And why can't I use the CMS's editor to fix the borkage?"
Me: "Where did you get the original text from?"
Content editor: "I copied it from a Word doc that somebody sent me. I just pasted that in. It was just plain text..."
Me: "I see. Well, delete the page and start again. This time, copy the stuff from Word, then open Notepad, past the text from Word into Notepad, then copy/paste into the CMS from there instead."
Content editor: "Oooh, voodoo!"
Me: "Indeed."
MS-DOS 7.0 workaround (Score:5, Funny)
I remember when Microsoft put a crappy 32-bit front-end on MS-DOS 7.0 to make it more useful. It completely sucked. It hogged memory and crashed all the time. Luckily you could boot directly into DOS to avoid the GUI and get real work done.
Not excactly a workaround (Score:5, Funny)
I used to have a network with windows NT 3.51 box and several 95 workstations.
Several times an hour I would see on the NT box a log error saying "An unexpected error has occurred on virtual circuit X."
NT 3.51 came with an online ref book you could use to look up things like that. When looking up the error code the page only said something like:
"If you expected this error ignore it."
Customer Service App (Score:5, Funny)
Just thought of another one:
Many years back I was working as a freelancer developing the training material for a customer service app.
The agents input customer details, the app identified the nearest call-out contractor, sent the contractor a text message, started the clock ticking and updated the log.
Unfortunately, the devs used their own GUI and in the top row the 'submit' button was right next to 'form clear' and call centre staff kept clicking the wrong button, erasing the customer details and having to ask for them all again. This did not go down well with customers who'd called due to a domestic emergency (plumbing etc.)
I suggested that the workflow through the form meant that the agents would be better served by a submit button at the bottom.
The response to my submission: "Can't see a need to move the button during this development cycle - agents to be told to stop clicking the wrong button."
not really a bug but (Score:5, Funny)
Qqest GoldSuite Timeclock Software ... (Score:5, Interesting)
So, if you had purchased 15 licenses and were running 2 or more clocks (but less than the 15 you were supposedly allowed), you'd run out of licenses after a couple of days, even with light use.
After working for a month or so with the company to resolve the issue, what was their long term solution?
Give us a code that would give us "unlimited" (or somewhere in the area of 32,000 licenses).
After several years (like 8 or so) and much griping from me to switch to something else, we're still using the software, actually (but with only one swipe station, and only for our student workers in our biggest department), but will supposedly switch to something hosted and web based "soon".
The case of the 500-mile email (Score:5, Interesting)
Ubisoft DRM fix (Score:5, Interesting)
Apple Mac only CD Rom (Score:5, Funny)
We labeled 3000 free handout CD roms "Apple Mac only" when we discovered that there was a windows virus on all of them. Clever huh?
Atari + tape recorder (Score:5, Interesting)
These weren't official ones. I developed them on my own.
The tape recorder was notoriously difficult to get the data to load right. Some tapes, saved on a different recorder, would require special tricks to get the readout "within specs".
One, I had to mute audio in the TV set to which the Atari was hooked up. I guess electromagnetic interference from the speaker was a problem. :)
On another, I'd have to hold the label with key functions on the recorder. The label was metal and connected to the recorder ground. By holding it, I was providing extra grounding that reduced the noise just enough to get the game to load. Luckily that one took only like 5 minutes to load
The best one was copied from a floppy. The copy was good, but there was no 'loader' program and the game was too big to fit with a copier to copy it to a different tape, and recorded from the beginning of the tape, no room to save the loader. The solution was to take a random different tape with a generic loader, start loading it, then after counting 6 "beeps" QUICKLY remove it and put the right tape in - the timeout tolerance was like 2-3s, so you really had to hurry.
HIT your Sun workstation (Score:5, Interesting)
Way back when dinosaurs roamed the earth, ok not really, but in the late 80s, Sun had a problem with some of their hard drives. When they would park they heads they would stick and you couldn't unpark them. Sun's solution was to tell you to HIT the computer. They even sent us a letter showing you where on the "pizza box" enclosure one would strike.
two second 'nop' (Score:5, Funny)
After upgrading a server, we watched a client verify the server through his daily application. The client entered data and clicked on submit, the next screen appeared instantly. "This is not possible" said the client "it takes about two seconds to submit data to the database"!
"But the new server is much faster!" we said. It didn't matter, the client refused to believe the data was really submitted.
We held a meeting about this 'problem'. One developer suggested to add a two second 'do nothing' loop to the submit button.
So we patched the server and asked the client to verify again. He entered data, clicked 'submit' and was very happy to have his two second delay back! "Now it works..." he said "...now the data is entering the database!".
We admitted our fault (knowing very well that all we added was a two second delay).
cheers
It's not a bug... (Score:5, Interesting)
I used to manage Digital UNIX (later called Tru64) systems for a large, now bankrupt, telecom back around the turn of the millennium. The filesystem used, AdvFS, was pretty cool and advanced for the time but under the version of the OS we were running we found that free space would shrink at a faster rate than used space would grow. I had filesystems report full even though a df would show only 60% used.
It turned out that when small files were deleted all of the space wouldn't become free. My customer wrote thousands upon thousands of 150-200 byte files a day and deleted just as many. The entire team and my customer agreed this was clearly a bug.
When brought up with Compaq (who had recently aquired Digital) the technical rep investigated and reported "this is not a bug, the code is being executed exactly how it's written." Seriously, this was his response. I would have been more amused if he seriously argued it was a "feature."
I never could get a definition of what a "bug" really was from him. I became rather infuriated when he reported to me that this issue was "fixed" in the latest major release of the OS. If there was no bug, why was it fixed?
I never got a straight answer and was left on my own to find my own work-around which involved inserting a new volume into the filesystem thus growing it and then deleting an old volume. When this was done to all volumes in the filesystem, the problem was resolved for a few more months. This was an incredibly labor intensive and, as far as I'm concerned, incredibly risky to move data around like that on a hot system with insane uptime requirements. There was also a massive performance hit while this was happening and my customer's application was already VERY IO intensive.
I'm still just as angry about that conversation with the rep today as I was back then.
Re:rigoddamndiculous ? (Score:5, Informative)
John Wayne made it up:
http://www.celebrityrants.com/premium/celeb_wayne.html
Parent
Re:rigoddamndiculous ? (Score:5, Funny)
[...] self documenting and shouldn't have a definition [...] fan-fucking-tastic for example.
I understand what 'fan-fucking' means and 'tastic' is probably related to 'elastic' in some way, but the sexual perversities they invent these days...
Parent
Re:rigoddamndiculous ? (Score:5, Funny)
Point taken.
I will say that self documenting words (just like self documenting code) require a minimum intelligence level. I'm wondering what percentile of the US population you represented to get the "fan fucking" + "elastic" conclusion.
Parent
Re:rigoddamndiculous ? (Score:5, Funny)
I don't understand how people associate the word "fuck" so exclusively with sexual meaning. It seems to be a more prevalent attitude in America, affecting even supreme court justices.
Such a conclusion is a pretty unfair typecasting of such a versatile swearword. While "fucking" or "to fuck" is often used to describe sexual intercourse, the word has a great many other meanings. "Fuck off" being the most classic and familiar example, used to gruffly tell someone to remove themselves, or to desist from an action, etc, but perhaps only to express disbelief or some such. "What the fuck" shows the ability to use the word in an undirected fashion. Alone, "Fuck" can be an effective emotional outlet. "Fuckers" turns the verb into a noun, that is, if it were ever a verb in the first place. Things like "fan-fucking-tastic" show just how versatile this unique utterance can be, as it transcends classical descriptions.
So, "Fuck" is not just a sexual swearword. Perhaps, lacking any other terms, American's take it to primarily refer to intercourse. In fact, other english speakers have many other words at their disposal for describing sexual activities. "Shag","ride", etc. Lack of such words in someones personal or cultural lexicon should not be used to imbue unwarranted meaning to a speakers words in some kind of reverse irony.
When Bono said "fucking brilliant" at the Golden Globes, it was clear to any reasonable person that he meant the word as an adjective to brilliant, not as a sexual reference. This is doubly clear to anyone from Ireland. Nevertheless the FCC claimed that the word had and "inherently has a sexual connotation", in any context. And worse, the US supreme court agreed with them.
As someone who has been told on countless occasions by friends, family and countrymen to "Fuck off", or some such like, I'm personally offended far more by the suggestion that all these people's comments had an underlying sexual meaning than I am by any of the expletives themselves. But once again I find my culture, my traditions, my airwaves, and my internets subjected to the interpretations and censorship of conservative bible bashers in rural America. It's fairly insulting.
So please accept my sincerity when I say that you, and all those that would corral honest swearwords into narrow definitions, respectfully, Can All Fuck off with Yourselves!
Parent
Re:rigoddamndiculous ? (Score:5, Funny)
urban dictionary = idiots making up words.
At 27 years old I am now an old fart.
UUuuh hello??! Rigoddamndiculous is a perfectly cromulent word!
Parent
Re:wiggle their mouse continuously (Score:5, Informative)
IIRC, a few GNU encryption programs do the same thing while collecting entropy, and yell at you if you don't wiggle enough.
Feature. Not a bug.
Do you have any idea how hard random data is to collect?
Parent
Randomness at 96 kbps from an ADC (Score:5, Informative)
Do you have any idea how hard random data is to collect?
If your PC has a sound card, an entropy gathering service can hash the microphone input and derive at least 1 high-quality random bit per sample from ADC dither noise alone. So that's 96 kbps for a typical 48 kHz stereo ADC.
Parent
Re:wiggle their mouse continuously (Score:5, Informative)
I think this actually had a good reason.
A nice old PS2 mouse generates interrupts when wiggled. This breaks up the boring routines. (Blocking routines actually.) And presto, a little more progress on transfering your data...
This phenomenon is not gone btw.
1. Start notepad in a window, not full screen.
2. Open long text file
3. Mark your text from beginning of document and try to scroll down. When mouse exits window, keep holding but with mouse stationary. Nothing happens?
4. Wiggle mouse outside window and presto it continoues to mark text towards the bottom of your document!!!
Fun and entertainment for the whole family!
Parent
Re:Ok, (Score:5, Funny)
Parent
Re:HP Printers and Windows are a No Go (Score:5, Insightful)
speaking of HP printers, especially the networked ones, why is it that the network driver is 350 megs in size? I had to download two of those damn things, even after using a custom install option, to remove as much of the cruft as possible I still installed some 700 megs of drivers for two printers, and a scanner.
Guess what happens when the drivers get corrupted. you have to manually uninstall the registry settings and deleted all files manually in order to reinstall the drivers or they won't work.
HP decent printers, Software coded by monkey banging on keyboards.
Parent