It wasn't quite that bad - somebody wrote the mathematical programming optimisation bit in Fortran. Everything else, from editing schedules to calculating inventory profiles, was in VB and was very slow.
The worst thing was, the company's next product (sold for many, many thousands of pounds) actually used VBA for a major part of the interface, which meant the C++/Fortran developers wasted months dealing with bugs and interfacing problems with VB because management wanted a VB interface.
My first paid programming work was to write an off-line chemical process schedule optimisation and inventory management application in VB6/VBA. I still shudder everytime I think about it.
Why would anyone miss that language? Let alone bother to sign a petition to save it. If your job relies entirely on a language that your average 12 year old can pick up in a week or two, you're in trouble.
I have enough trouble explaining what I do to most adults, let alone small children. "Software developer? Is that like a photocopier repair man?" was probably my favourite response.
Crashing Nautilus on start up is pretty easy - just log in;). I'd say it happened about 25% of the time on FC3 - the bug reports in bugzilla from other people were equally vague, but it seems to have been fixed anyway. Crashing Nautilus in normal usage was pretty easy: just open two directory windows and drag files between them for a bit, and eventually Nautilus would die. They've only just got round to thinking about writing a test suite for browsing, so I'm not surprised it was unstable.
The last panel crash I can remember was after deleting a drawer containing a weather applet from a panel - that was in 2.10. I can't remember the exact causes of many other crashes, on FC3 they were so common they all blurred into each other.
I deleted my bookmark to osnews a while back after a previous one of these terrible articles, but ended up reinstating it after a few links from slashdot stories convinced me things had changed.
Never again: my osnews bookmark is gone for good. I guess I should have realised how bad a site it was when I sent a quick, badly written, email about an article to her and it was used verbatim as a story.
Alan Cooper is a well known UI expert. His comments on dealing with users are fairly similar to yours: design for the user, observe the user, but don't let the user do the design.
planet.gnome.org has a load of GNOME developers responding to the two articles in a far more logical and intelligent way than the articles deserve.
Somebody like Eugenia who runs such a badly-implemented news+comment site really shouldn't complain about GNOME not implementing features the users want.
I'd always assumed that there was a libGecko, libXUL , libMozMail etc. that Seamonkey, Thunderbird and Firefox all used to avoid duplicating effort. That doesn't seem to be the case, however (from my understanding now, anyway): they all seem to use their own, slightly different components.
Wouldn't it be a good idea to just split out as much of the common functionality between the various Mozillas. The "Firefox team" becomes basically a browser UI team, for example.
That way, if the libGecko people need to release a security patch, they don't need to wait for the firefox people to merge it into their source tree as well. Or is my understanding incorrect?
Before you wipe that Warty install, you might want to give the Hoary development release a chance. I switched to FC3 from Warty after getting pissed off with all the bugs, but FC3 wasn't much better. Hoary is currently far more solid than FC3 ever was.
I prefer GNOME as a desktop, but it has had some stability issues over the years. I think KDE benefitted from some developers pushing test-driven development, giving them a more solid base.
All companies do what is best for them, not the customer; you don't see companies selecting prices so as to maximise customer happiness, they set prices to maximise profit.
It just happens that competition tends to push companies so that what is best for the company is at least similar to what is best for the consumer. I don't know why the even bother saying they do what is best for the customer: why don't they become a non-profit if they care about the customers so much?
I'm not saying that profits are bad, just that I dislike this kind of doublespeak.
> Why would someone click on a menu item and hold?
That's what I do mostly - click on menu, move mouse to item and release. It works fine in GNOME, but sometimes with the applications menu it gets confused and thinks you want to drag a menu item off the menu.
Anyone have a decent benchmarks of Apple hardware for development related stuff - compiling etc. The only benchmarks I can find are photoshop and the dodgy spec benchmarks.
This is from "criminal law 4th edition", Michael J. Allen, 1997 (potentially things have changed since then):
"Section 4(1) defines property to be... things in action and other intangible property"
"A 'thing in action' is a personal right of property which can only be claimed or enforced by legal action and not by taking physical possesion... Examples of things in action are debts..., a copyright or trade mark..."
"the Patents Act 1977 treats an invention for which no patent has been granted or applied for as intangible property"
I was wrong in my analysis, of course: I missed the "in circumstances making it equivalent to an outright taking or disposal" from the end of my last point, but it's interesting anyway.
Maybe. Interestingly, looking at my girlfriend's UK criminal law textbook, you can actually be convicted of theft for 'stealing' intangible property under the Theft act.
Apparently things like patents, inventions and copyrights can be 'appropriated' (stolen), and "assumption of the rights of an owner" is counted as appropriation. Also "a person appropriating property belonging to another without meaning the other permanently to lose the thing itself" is counted as depriving the person permanently of it if "his intention is to treat the thing as his own to dispose of regardless of the other's rights".
I guess that's why copyright infringement is a criminal offence over here. Not that I understand much of the textbook of course.
It's certainly theft in the general use of the word: if I made a copy of an author's manuscript and published it under my own name, most people would be happy with calling that "stealing the author's work", although it's actually just copyright infringment.
Theft means taking without permission. The GPL only grants permission to 'take' the licensed source code if you obey certain restrictions. This guy doesn't appear to have met those restrictions, so he has stolen the code.
The point of the GPL, in case you missed it, is that modifications to the source cannot be kept from the community if the modifier wants to distribute their work. If you want to benefit from GPL code, you have to give back in the form of your modifications.
Isn't that attitude the reason that so many games are buggy pieces of crap with half the promised features missing?
Unless you're writing VBA for a living, software development is a mentally challenging process. It's physically not possible to keep up the same rate of coding for 20 hours. Your brain needs time to relax and process data.
Personally, I tend to come up with better solutions to problems a few hours after I stop working on them and my subconcious has had time to process.
Crunch coding is not only a bad idea for the workers, it's bad for the company too.
I've faced this problem too with checking visual output. What I will probably do at some point is do automated screenshot comparison: have the system do the test, then compare the relevant region of the screen to a known-good image as a regression test. The only problem I can see with that is that generating the known-good images is time-consuming and minor changes would require regenerating them all.
No, I'm saying that it is not helpful to submit unspecific bug reports like "it crashed when I opened lots of tabs". Developers have limited time - that's what this article is about - and anything you can do to help them is welcome. Bad bug reports waste the developers' time for no benefit; detailed, specific bug reports can save them hours.
If you are expecting them to fix a bug you have found, then the least you can do is detail the actions you took that caused the problem in as much detail as possible.
But I haven't presented an explanation. Am I talking to a bot here or something? All I said was you didn't provide the information necessary to determine an explanation. If you were serious about getting the bug fixed, you would start removing variables and see if it still reoccurs. There are too many possible causes for your bugs. 98% CPU, for example, could be caused by javascript gone haywire, java applets, flash, an mplayer-plugin, and is not necessarily connected to the memory issue.
You don't give up. The point is, your list of problems is basically useless from a bug-hunting point of view: it provides no information to help the developers at all
If you had reported these as seperate bugs, with some indication of URLs that were open at the time and without all the "firefox is mission critical and godly" stuff, you probably would have got a better response. Just accept that and remember next time you report a bug.
As it stands you have said "there's a memory leak somewhere in the code". If you phoned your local highway maintenance people and said "there's a street light broken in London", you wouldn't expect to get much of a response. That's what you have done.
What you see as anger on the part of the developers is most likely frustration at your inability to understand the requirements for reporting a bug.
Slashdot Mirror
It wasn't quite that bad - somebody wrote the mathematical programming optimisation bit in Fortran. Everything else, from editing schedules to calculating inventory profiles, was in VB and was very slow.
The worst thing was, the company's next product (sold for many, many thousands of pounds) actually used VBA for a major part of the interface, which meant the C++/Fortran developers wasted months dealing with bugs and interfacing problems with VB because management wanted a VB interface.
Hope they don't read slashdot, anyway.
My first paid programming work was to write an off-line chemical process schedule optimisation and inventory management application in VB6/VBA. I still shudder everytime I think about it.
Why would anyone miss that language? Let alone bother to sign a petition to save it. If your job relies entirely on a language that your average 12 year old can pick up in a week or two, you're in trouble.
I have enough trouble explaining what I do to most adults, let alone small children. "Software developer? Is that like a photocopier repair man?" was probably my favourite response.
Crashing Nautilus on start up is pretty easy - just log in ;). I'd say it happened about 25% of the time on FC3 - the bug reports in bugzilla from other people were equally vague, but it seems to have been fixed anyway. Crashing Nautilus in normal usage was pretty easy: just open two directory windows and drag files between them for a bit, and eventually Nautilus would die. They've only just got round to thinking about writing a test suite for browsing, so I'm not surprised it was unstable.
The last panel crash I can remember was after deleting a drawer containing a weather applet from a panel - that was in 2.10. I can't remember the exact causes of many other crashes, on FC3 they were so common they all blurred into each other.
I deleted my bookmark to osnews a while back after a previous one of these terrible articles, but ended up reinstating it after a few links from slashdot stories convinced me things had changed.
Never again: my osnews bookmark is gone for good. I guess I should have realised how bad a site it was when I sent a quick, badly written, email about an article to her and it was used verbatim as a story.
Alan Cooper is a well known UI expert. His comments on dealing with users are fairly similar to yours: design for the user, observe the user, but don't let the user do the design.
I've seen several panel crashes and lockups with GNOME 2.8, plus many, many Nautilus crashes. Nautilus crashing on startup became the norm on FC3.
Most of that seems to be fixed now with 2.10: it was a major bugfixing release.
planet.gnome.org has a load of GNOME developers responding to the two articles in a far more logical and intelligent way than the articles deserve.
Somebody like Eugenia who runs such a badly-implemented news+comment site really shouldn't complain about GNOME not implementing features the users want.
I'd always assumed that there was a libGecko, libXUL , libMozMail etc. that Seamonkey, Thunderbird and Firefox all used to avoid duplicating effort. That doesn't seem to be the case, however (from my understanding now, anyway): they all seem to use their own, slightly different components.
Wouldn't it be a good idea to just split out as much of the common functionality between the various Mozillas. The "Firefox team" becomes basically a browser UI team, for example.
That way, if the libGecko people need to release a security patch, they don't need to wait for the firefox people to merge it into their source tree as well. Or is my understanding incorrect?
Before you wipe that Warty install, you might want to give the Hoary development release a chance. I switched to FC3 from Warty after getting pissed off with all the bugs, but FC3 wasn't much better. Hoary is currently far more solid than FC3 ever was.
I prefer GNOME as a desktop, but it has had some stability issues over the years. I think KDE benefitted from some developers pushing test-driven development, giving them a more solid base.
All companies do what is best for them, not the customer; you don't see companies selecting prices so as to maximise customer happiness, they set prices to maximise profit.
It just happens that competition tends to push companies so that what is best for the company is at least similar to what is best for the consumer. I don't know why the even bother saying they do what is best for the customer: why don't they become a non-profit if they care about the customers so much?
I'm not saying that profits are bad, just that I dislike this kind of doublespeak.
2.10 is much better for this - I haven't had a panel or nautilus crash in the week of installing it, and that's on the pre-release of Ubuntu Hoary.
GNOME definitely needs some better testing procedures, I agree.
Somebody pointed that out the the artist already - the final version is meant to have that fixed.
> Why would someone click on a menu item and hold?
That's what I do mostly - click on menu, move mouse to item and release. It works fine in GNOME, but sometimes with the applications menu it gets confused and thinks you want to drag a menu item off the menu.
Anyone have a decent benchmarks of Apple hardware for development related stuff - compiling etc. The only benchmarks I can find are photoshop and the dodgy spec benchmarks.
This is from "criminal law 4th edition", Michael J. Allen, 1997 (potentially things have changed since then):
... things in action and other intangible property"
"Section 4(1) defines property to be
"A 'thing in action' is a personal right of property which can only be claimed or enforced by legal action and not by taking physical possesion... Examples of things in action are debts..., a copyright or trade mark..."
"the Patents Act 1977 treats an invention for which no patent has been granted or applied for as intangible property"
I was wrong in my analysis, of course: I missed the "in circumstances making it equivalent to an outright taking or disposal" from the end of my last point, but it's interesting anyway.
Maybe. Interestingly, looking at my girlfriend's UK criminal law textbook, you can actually be convicted of theft for 'stealing' intangible property under the Theft act.
Apparently things like patents, inventions and copyrights can be 'appropriated' (stolen), and "assumption of the rights of an owner" is counted as appropriation. Also "a person appropriating property belonging to another without meaning the other permanently to lose the thing itself" is counted as depriving the person permanently of it if "his intention is to treat the thing as his own to dispose of regardless of the other's rights".
I guess that's why copyright infringement is a criminal offence over here. Not that I understand much of the textbook of course.
It's certainly theft in the general use of the word: if I made a copy of an author's manuscript and published it under my own name, most people would be happy with calling that "stealing the author's work", although it's actually just copyright infringment.
Theft means taking without permission. The GPL only grants permission to 'take' the licensed source code if you obey certain restrictions. This guy doesn't appear to have met those restrictions, so he has stolen the code.
The point of the GPL, in case you missed it, is that modifications to the source cannot be kept from the community if the modifier wants to distribute their work. If you want to benefit from GPL code, you have to give back in the form of your modifications.
Isn't that attitude the reason that so many games are buggy pieces of crap with half the promised features missing?
Unless you're writing VBA for a living, software development is a mentally challenging process. It's physically not possible to keep up the same rate of coding for 20 hours. Your brain needs time to relax and process data.
Personally, I tend to come up with better solutions to problems a few hours after I stop working on them and my subconcious has had time to process.
Crunch coding is not only a bad idea for the workers, it's bad for the company too.
I've faced this problem too with checking visual output. What I will probably do at some point is do automated screenshot comparison: have the system do the test, then compare the relevant region of the screen to a known-good image as a regression test. The only problem I can see with that is that generating the known-good images is time-consuming and minor changes would require regenerating them all.
No, I'm saying that it is not helpful to submit unspecific bug reports like "it crashed when I opened lots of tabs". Developers have limited time - that's what this article is about - and anything you can do to help them is welcome. Bad bug reports waste the developers' time for no benefit; detailed, specific bug reports can save them hours.
If you are expecting them to fix a bug you have found, then the least you can do is detail the actions you took that caused the problem in as much detail as possible.
But I haven't presented an explanation. Am I talking to a bot here or something? All I said was you didn't provide the information necessary to determine an explanation.
If you were serious about getting the bug fixed, you would start removing variables and see if it still reoccurs. There are too many possible causes for your bugs. 98% CPU, for example, could be caused by javascript gone haywire, java applets, flash, an mplayer-plugin, and is not necessarily connected to the memory issue.
You don't give up. The point is, your list of problems is basically useless from a bug-hunting point of view: it provides no information to help the developers at all
If you had reported these as seperate bugs, with some indication of URLs that were open at the time and without all the "firefox is mission critical and godly" stuff, you probably would have got a better response. Just accept that and remember next time you report a bug.
As it stands you have said "there's a memory leak somewhere in the code". If you phoned your local highway maintenance people and said "there's a street light broken in London", you wouldn't expect to get much of a response. That's what you have done.
What you see as anger on the part of the developers is most likely frustration at your inability to understand the requirements for reporting a bug.