Not sure that reinstall fixed it, maybe user didn't commit the same error, whatever it might have been.
certainly possible, but not probable when multiple experienced people see this problem and its repetable and with no ideas about why this should happen. 80% of all bugs may be caused by ignorant users. But there are plenty of people who find real bugs from part time jocks to full time professionals. But there are still millions and millions of bugs everywhere. As you incrase your knowledge of computer archectecture and writing software you learn first hand about all the pitfuls that lie in a programmers way. Mistakes happen constantly. Despite top down design, detailed functional specifications, Object oriented progamming and I still write code all the time with bugs. I always test the code at multiple levels using several techniques to verify functionality subjectively. Additionally objective measures of software engineering can be applied including code coverage including block, path, and expression coverage,code profiling and run-time memory heap analysis. These provide ways to help increase your reliability and chances for writing robust code. I deal routinely with CAD tool vendors who sell expensive software for $20K to 100K+ per individual license. Then they hit you up 15-18% of purchase price a year for maintenance which includes TS and minor upgrades. You would think that they of all people would write great code. unfortunately not, 1 1/2 years ago I was dealing with a new simulator which simulated VHDL and Verilog with a single kernal. During my first two months I logged over 40+ problems. In 4 months I had amassed 65 problems with their product incluing 14 Severity 1. Their development team was understaffed and couldn't seem to write solid code. Their president, vp sales, vp ts, flew out to visit us in Raleigh as it escalated and brought me 180+ Page Notebook of all my problem logs and then showed me some statistics. Normally they received aproximately 200 ts questions a month. of which only 7% result in the creation of a software defect report. I had logged 82 calls in to them of which 65 had resulted in defect reports. They were stunned and didn't know what to do. I had found more bugs in their code in 4 months than there whole customer base combined. Needless to say this relationship didn't go real far. I find lots of other bugs in lots of tools all the time and make it point to promptly reported routinely repeatable or severe bugs as soon as possible. Its a good habit others should get into. At least another 10 were bugs that they were too lazy to fit. I logged many bugs to gary of which most he will tell you were bugs. Just recently I have logged over 10 confirmed bugs with Mijenix on PowerDesk Explorer 3.0, NT's DHCP service is a pandora's box of bugs... I could go on for hours...
Yes, we might find more features (bugs)like the black cursor. I had heard of it but never had a desire to use the cross-hairs. At least QP continues to evolve, and provide updates. So many programs turn out a version and leave us hanging until they decide to try the next version
absolutely, gary is trying real hard to deliver a superior product and like everyone else is limited by time and money... He needs a TS call tracking system and bug tracking system to improve his ability to track this stuff. I can't imagine writing large peices of code without these tools. you end up with too many loose ends....
Even Gates can be shown to offer free fixes and features. Glad I don't have to write programs, manuals, or help systems. You can't win for losing.
MS's concept of fixing bugs is hardly exemplary. Lets compare some other computing platforms. Two I'm quite familar with are IBM RS/6000 running AIX, Sun SPARC running Solaris. IBM AIX Version 3 has near 3,000 patches avaialble for it in several years. Sun Solaris 2 has nearly 5K patches avaialable for it as of last week. Both these OS's have functions which keep track of all installed patches and system library versions. You can go through hypertext and select which ones you want and download them or they will burn you a CDR and send them to you for installation. Microsoft might have released 100 bug fixes/enchancements to 95 in two years. You would think windows is more stable and hass les bugs but the sad truth is the total opposite. I use all of these everday in my work and at home for about 8 years now. There is no comparison unix is much more stable by at least a factor of 10 without these fixes. With all the latest patches on AIX or Solaris you are looking a extremely stable platform that will generally run for months or years without crashing. My Latest SUN Ultra has not crashed once now in 18 Months. My NT boxes crashes maybe once every six weeks but suffers other problems. 95 is hopeless for hard core use. daily crashes or hangs. The point being that there are thousands of bugs that are never documented is MS KB or elsewhere and there are many in the KD that are dated quite old which have never been resolved. The basic unstability of 3.1/95 is due to poor archectecture on MS's part. The bugs are generally caused apps and Device Drivers who perform illegal operations (too many to list) and windows can't protect against it. These flaws exist in order to maintain windows 95 backward comaptibility with dos and speed. Modern Robust OS's all have a tight secure kernal which interface to the hardware and all everything else has to talk to the kernal. By addding this abstraction the kernal can detect exceptions and IMMEDIATELY kill applications that are performing illegal operations and protect any other programs code or data running on your system for being corrupted. The result is much more stable robust platform. The sad thing is all the cool personal software runs on windows more and more often. IBM,Microsoft, Intel have succeeded in making pc's and software commonplace and affordable to the masses which no vendors had done previously. Many software engineers fail to test their software adequately or take adavantage of such tremenously powerful tools such as heap debuggers coupled with MS's poor inter-program protection results in windows 95/98 instability. NT provides hope but still has some nasty holes that cause far too many blue screen exceptions. Thats another subject. BTW: Unix doesn't have the problem with all the DLL's landing in windows sytem and overwriting each other. Each program will used registered system copies or include their own copies in the program directory. There is no registry either. Everything is stored in configuration files in the users directories or in the program directory so you only have to install a program once and can run it on 10,000 Clients. Truely network friendly. MS office only dreams of such network robustness...
How can you call software network friendly that requires client side installation on each machine and turns all images into HUGE BMP's rather than jpegs or gifs... boggles the mind :)
I know many of you grow accustomed to having worrying about crashes or rebooting often, or even installing some new software that causes your system to behave abnormally. you may think your doing something wrong, you probably are, but not necessarily The sad truth is that it doesn't have to be this way, I know because I do use computers without these problems everyday. Windows is the only modern OS I can think of that still lacks solutions to these problems.....
Sean |
