ZD magazines have some credibility compared to some of the junk in the UK.
Linux Format must be the most pathetic excuse for a magazine I've seen since the 8-bit explosion of 1983. Five pounds for lots of irrelevant pictures, large display fonts and inaccurate content.
I've got a bet on that it won't last beyond the end of this year.
The BBC is funded by my television license, which costs around 100UKP per year. This pays for the majority of the BBCs output, 2 terrestial television channels, 3-4 additional digital channels, 5 national radio stations, regional radio for Scotland, Wales and Northern Ireland and 30+ local radio stations for England. The world service is separatly funded.
Yet another example of dumbing down to meet the great unwashed.
I used to use Deja as part of my work, to quickly search through archives of a handful of specific newsgroups. I didn't mind the banner ads (advertising doesn't work on me), I could easily find what I wanted. I'd still use my normal account to post.
Then they introduced the first changes. Finding what I wanted was a pain, but still useable.
Now they've added this junk, aimed to target the point-click-drool types even further. Particularly that the links don't really help
It's getting the same in the music press over here (the UK), with "If you like this, try this" sort of articles.
The proliferation of the jobs market today with MCSEs so clueless that they shouldn't be put in charge of a pocket calculator. has made the entire certification process a joke.
When I qualified for my CBE (Certified Banyan Engineer) 5 years ago it needed a lot of study, and the ability to think. You certainly couldn't pass this exam purely by regurgitating facts (or opinions, based on some of the MS sample questions I've seen). One exam consisted mainly of parsing sequences of protocol hex-dumps. Heck, most MCSEs I've met don't even know what hex is!
Nowadays anyone with enough money (or an employer with enough money), can obtain a certification.
I've done some reverse engineering, but in my case it was trivial.
I had an ISA sound card whose MIDI port was not working. The problem was that the chipset used by the card was of a later revision to the one supported by the driver in the then current Linux kernel. All I needed to do was to locate the IO port that contained the MIDI settings, then write the correct value to the card. No specs had ever been released, and the manufacturer no longer made sound card chipsets
I played around with the standard Linux kernel driver, adding printk statements to work out what was going wrong, and what values the probe routines returned. I worked out where in the code I could init the port, then got round to discovering what value(s) to write, and which port(s) to write them to.
I located the DOS driver that was loaded in Windows to initialise the card, without any documentation as to the switches to use. A quick inspection of the disassembly located the code that wrote from memory to the range of control I/O ports. I then located a table of IO Port value - bitmask pairs. Next to this table in memory was a simillar list of IRQ - bitmask pairs. By ANDing these pairs together I had the value to write to the IO register. I only needed the register itself.
I then searched the disassembly for a reference to the address of these tables - no such luck. Searching for addresses closer to this located the code that did the table lookup and wrote the correct value into memory. I hardcoded the write into my driver, reloaded the module, and all was now working.
I then spent a few minutes adding a couple of case statements and the odd debug message, then created a patch and submitted it. My patch appeared in the devel kernel a few weeks later.
Since then I've acquired a 6-port serial/synchronous card. Not only do no documents exist for this card, but there only code I can locate is a DOS diagnostic. The card was expected to run under a custom OS which I don't own. I'm not having much luck.
With binary only drivers, you are usually only tied to one particular kernel on the i386. uniprocessor platform. Need a driver for an SMP platform? What about an Alpha?
You are often tied to the specific config of the machine the driver was built on (usually the current RedHat). How many times have you run make config/menuconfig under 2.2.x, changed a single parameter, and discovered that half the kernel source needs to be recompiled?
The other problem with binary-only drivers is that the source has not been audited by one of the core kernel team (Linus, Alan or one of the others). You've no idea what the driver is doing. Is it susceptible to races? Does it tie up CPU time too much? Does it try to probe the hardware in a bad way? Does it use depreciated calls? Is there a memory leak somewhere? And so on. I've had enough of problems like this with add-on drivers for other Unix variants in the past to be very wary of 3rd party drivers.
My former company was into network support and software reselling, based in the UK. I once saw a presentation that listed customers. It included a couple of big companies, both of which had boughty the same piece of software from us in the past, 2-3 years ago, but no longer used it, as the manufacturers had abandoned all support and development.
They later sold a custom web server product without telling anyone that the backend was Lotus Domino.
The only problem I've seen with the current version (0.49 - released last month) was when resizing when connected to one particular Solaris 2.5 system over a slow and not too reliable connection. Occasionally the terminal session would still have the previous dimensions. Resizing when connected to an indentically configured machine, located on the same network segment as my PC, works fine.
I currently have it configured to use Andale Mono, one of the extra fonts bundled with IE5.0. The beauty with this font is that even at small point sizes all characters are distinct.
Whilst adding all the DHTML and other gubbins to IE5.5, Micros~1 still havn't bothered to get PNG rendering working correctly, as can be seen here. The output is exactly the same witn IE 5.0 and IE 5.5.
Oddly, IE5.0 on the Mac renders PNGs correctly.
Typical greedy bloody Micros~1, always adding new features to embrace and extend, without fixing things that are known to currently be broken.
I'd really like to remove all the IDE Bus-mastering code, and have the processor use PIO for everything.
Can I have a pint of what he's drinking.
on
X Windows Must Die!
·
· Score: 3
Is he drunk or something??
Seriously, it seems that he doesn't understand the underlying flexibility of X - that the display and application can be in separate locations.
This is something that the Windows world lacks. So PC-Anywhere et al allow remote control, but these are limited to one simultaneous controlling user. (Ever tried to remotely connect to a NT box to run some admin utility that can only be controlled from the console, only to find that someone else is controlling some other application at the same time).
A window manager is a matter of choice - you don't even need one if only running one application under X. I currently run fvwm2, mainly to the small footprint and a cronic lack of disk space. I could be running sawmill/sawfish this weekend (he didn't mention that one), or maybe try the latest CVS snapshot of enlightenment.
His other main criticism was on the size of the code. The QNX installation he quoted is a very basic environment, I believe only using basic VESA modes with no acceleration. Yes it fits on a floppy, but then I ran Gem/DOS 3.2 from a single 360K floppy 11 years ago. X (and the Linux kernel itself) is certainly more complex in nature than the QNX demo software (which is only a demo after all, and not a fully featured OS).
Oh, and he got the name wrong - there's no such thing as X Windows. I have an instant distrust of journalists who get the name of software wrong - it never bodes well on the content of their articles.
I spent 2 days (that was enough) working for a UK hardware vendor.
I am certain they fiddled their reviews. Kit sent out for review was built on a different production line to a higher standard (eg cables were tied up, instead of left to hang loose). The casing was usually of a higher quality, and the parts generally top range. Paying customers (particularly the genral public), got whatever junk could be piled together in a day - often hardware that had been used elsewhere.
I am also sure I saw one of the 'technicians' activly running benchmarks for one machine in the test-lab, and these benchmarks were to be provided to one of the magazines. Oddly, this company usually scored very well in the benchmarks - I wonder why. (Personally, I regard all benchmarks with disdain).
I walked out of the place after 2 days. Not only was I disgusted by the way the company worked, but they treated staff like dirt.
Look what happened to Banyan.
on
Endgame For SCO
·
· Score: 1
Banyan were the other major player in the NOS markets for years, particularly when it came to large organisations. Now where are they?
The VINES os was based upon AT&Ts UNIX of the late 1980s, although with some modifications to the tasking mecahnism, and later the filesystem. Ports were made to SCO, AIX, HPUX, Solaris and other UNIX flavours (SINIX?), but due to a combination of bad marketing (always Banyan's weakness) and the fact that the UNIX port was based upon the features of the previous version of the native OS.
If configured properly, it worked well, and was used by many big organisations (Intel, Compaq, USMC). Some of these networks had around 1000 fileservers, all interconnected.
Banyan had a number of problems which contributed to their downfall. Hardware for their own proprietory servers (discontinued around 1991) was expensive. Marketing, as mentioned earlier, didn't exist. A port of their StreetTalk directory services was released when it clearly wasn't ready. Third parties were not able to develop device drivers until the mid 1990s (it took almost a year for a working server 3c5x9 driver to appear). Oh, and marketing from Redmond played a factor too.
One major contributing factor was the mail client. The DOS one worked fine, and was ported to OS/2 and Unix (the latter only working for shell accounts on a system also running as a Banyan server). Instead of developing their own simple one, Banyan bought out Beyond Mail. This mail system had rules processing, but these were never integrated into the process running on the mail server, and needed its own engine. The first Banyan labelled version of Beyond worked (sort of), but the next one was a mess. Slow, buggy, always crashing with 'assertion failed' errors, and not able to cope with the year 2000. There was one 3rd party mail client available that was still maintained - sites either switched to that or abandoned Banyan's email. Once Email was running on another system the integration of users as an object was lost.
The last straw was for Banyan to announce an agreement with Microsoft, basically stating that they would migrate all current users to NT. This too lost a lot of customers.
Banyan still just about exist, although only legacy customers remain. Most sales that I was aware of during the last 4 years were upgrades, with a few new servers to existing customers. The major players (Compaq, Intel etc) swiched over to NT (spit!)
If Banyan has invested some of their UNIX IP into offering a Linux port, they may still be going concern. They did announce an 'open' version of StreetTalk a few years ago, but I don't recall anyone actually getting hold of it. Some of the aspects of their protocols were very closed, particularly security, so a OSS clone would be difficult. Once OSS cloning was mentioned on the Banyan mailing list most customers had already given in the towel.
I supported SCO Xenix and Unix systems in the past. Support was poor. Known problems were not documented. Hardware support was difficult to get hold of. The usual stories with non mainstream commercial software.
I became the only person in the UK supporting Banyan's VINES on SCO Unix, and had real problems getting support from anyone knowledgable through the official channels. I felt things could be better. and had to learn everything by trial and error.
In 1995 I was tasked with setting up our DNS server on an existing SCO system which ran other tasks. The BIND implemenation shipped with the current SCO UNIX TCP/IP implemenation was ancient even then, and just wouldn't work with our then ISPs DNS. A few days of attempting to get help from SCO got me nowhere, so I grabbed a PC destined for use as a print server, upped the RAM slightly, snarfed Slackware via FTP, and had a working DNS server within a couple of hours, which not only ran quicker, but also had all the bits and pieces that cost SCO user an arm and a leg (TCP/IP, NFS, C compiler, nroff and so on). I started to relalise the SCO was overpriced junk.
We were still an official reseller/integrator/whatever, and still receive the glossy magazines, although they didn't come to me:-(. 3-4 years ago I finally saw one on a sales-droid's desk. Despite SCO being the magazine's title, over 75% of the magazine content was NT - and this was an official publication. I knew SCO had troubles.
I'm just suprised that they've lasted so long. Only the actions of suits in preferring SCO over Linux or *BSD ('We need someone to blame when things go wrong! - Waddya mean no-one;s sued Microsoft') has kept them afloat for so long.
I want them to either go out of business, or to suffer the ultimate fate bestowed upon the dead - a buyout from Computer Associates.
I'll wager that the response time is due to a certain organisation in Redmond has released their new web browser within the last 24 hours, and every looser on the planet is currently downloading the thing, soaking up bandwidth.
Performance here in the UK sucks badly at the moment.
There was a report on a UK television programme a few years back concerning a diasabled woman who turned up to vote, but was told that her proxy had voted on her behalf. In the UK a proxy can be nominated to vote on your, but a voter can still vote before their nominated proxy, if needed.
It appeared that one of the local parties had been registering themselves as proxies for a number of local disabled people, then voting on their behalf. I don't think any charges were ever made, which was odd as the majority for this seat was very close.
It is the most unstable piece of commercial I have to use at work. Generally my NT Workstation has uptimes measured in weeks, mainly due to me ensuring that all device drivers are current. However once I run JBuilder I often have to reboot. JBuilder seems to want to write directly to screen memory during startup.
If I can get hold of the debugger copy of Windows NT I'll have a look to see what it does do. I did do simillar once with Notes under Win 3.11, and was astounded by the number of incorrect calls that POS was making.
I've come across even worse security problems with commercial software. The following represents a single incident out of many.
The organisation I used to work for had a department that decided to sell a commercial Firewall (no longer in production, thank $DEITY). A copy was installed in our Office to see if it worked. The clueless fool who installed the thing didn't understand how FTP worked, and we were without FTP from a browser until my department got the firewall's password, when suddenly everything worked.
My department didn't have anything to do with the configuration at customer sites, but were called if things went wrong. The pilot site called a few weeks later to say that had been implicated as a source of spam. Of course, my team knew nothing about the config of this system, but were still expected to resolve it.
Digging around revealed that the SMTP process on this firewall was being used as an SMTP relay. We contacted the UK distributer of this product. We eventually got a replay, stating that the fix would be available in the next maintenance release of the firewall, currently scheduled to be released in 6 months time. Given that this was a major security hole, the fix should have been out as soon as the hole was discovered.
I don't recall the fix ever being released. I do recall around a year later a collegue stripping down all PCs in the building, looking for older versions of the 3C509 NIC to install in our filewall, as it was all that it still supported, so I'm wary if anything was fixed.
A while after that I received a call from a recruitment agency. The company responsible for the firewall had abandoned it, and now the UK distributers were now developing it themselves. They wanted someone to work in customer support. My response was two works, the second one being 'Off!'.
I've had enough of working in telephone support and testing, and need to find some means of getting into plugging security holes full time. Reading articles like this, and playing with tools such as Nessus has convinced me that there are a lot of exploitable platforms out there. My home machine is secure, the only possible hole being my X server which I'm going to fix tonight.
I've seen how lax many major organisations are when it comes to security (hey, we'll install a firewall, that will fix it), having worked with some sites that were a nightmare. Getting some customers to install ciritical patches was impossible until we refused to take any more telephone calls until they planned an upgrade.
Unfortunatly, there's not much scope in the security field at my current place of work (too much NT). Anyone with any ideas on how to progress - BTW I'm based in Yorkshire, England.
Intel had already released an ethernet card called the PC586 prior to the release of the Pentium. I always thought that they chose the name Pentium to avoid confusion over names.
BTW the PC586 was a pain to configure, compared to the then current 3Com cards.
The name 'Floopy' springs to mind. I saw them advertised in UK magazines for a few months around 1984. A bit bigger than a 3.5" floppy. Died pretty quickly. The name came from floppy and loop/
There was aslo something called the Hobbit drive, but I can't recall if this too used tape loops.
Are the data catridges used by Dec Vaxen loops? I saw one once, but didn't have the nerve to take it apart.
There has already been a storage system called a Microdrive, and they were hopeless.
In 1982 Sinclair annouced Microdrives for the ZX Spectrum, although it took a year for them to arrive. The QL and ICL One-Per-Desk (a QL clone with integrated Lan and telephone) also used them.
They consisted of a loop of tape in a matchbox sized case. Capacity was approx 80-100K (wow!). Reformatting after use would increase the capacity due to tape stretch.
They were unreliable anf propriatory (only Sinclair made them, and they sold initially at 5 UKP per cartridge IIRC). They were also slow due to seek times in seconds.
Typical Sinclair - designing their own solution when cheap storage, such a 3.5" floppies, was already emerging.
Slashdot Mirror
Linux Format must be the most pathetic excuse for a magazine I've seen since the 8-bit explosion of 1983. Five pounds for lots of irrelevant pictures, large display fonts and inaccurate content.
I've got a bet on that it won't last beyond the end of this year.
The BBC is funded by my television license, which costs around 100UKP per year. This pays for the majority of the BBCs output, 2 terrestial television channels, 3-4 additional digital channels, 5 national radio stations, regional radio for Scotland, Wales and Northern Ireland and 30+ local radio stations for England. The world service is separatly funded.
Not bad for around 2 pounds a week, is it?
I used to use Deja as part of my work, to quickly search through archives of a handful of specific newsgroups. I didn't mind the banner ads (advertising doesn't work on me), I could easily find what I wanted. I'd still use my normal account to post.
Then they introduced the first changes. Finding what I wanted was a pain, but still useable.
Now they've added this junk, aimed to target the point-click-drool types even further. Particularly that the links don't really help
It's getting the same in the music press over here (the UK), with "If you like this, try this" sort of articles.
Is it me, or is the world full of stupid people?
When I qualified for my CBE (Certified Banyan Engineer) 5 years ago it needed a lot of study, and the ability to think. You certainly couldn't pass this exam purely by regurgitating facts (or opinions, based on some of the MS sample questions I've seen). One exam consisted mainly of parsing sequences of protocol hex-dumps. Heck, most MCSEs I've met don't even know what hex is!
Nowadays anyone with enough money (or an employer with enough money), can obtain a certification.
Bill the Bastard has a lot to answer for.
Looks like I may be needing to look for work soon - oh joy, having to deal with these morons again....
I had an ISA sound card whose MIDI port was not working. The problem was that the chipset used by the card was of a later revision to the one supported by the driver in the then current Linux kernel. All I needed to do was to locate the IO port that contained the MIDI settings, then write the correct value to the card. No specs had ever been released, and the manufacturer no longer made sound card chipsets
I played around with the standard Linux kernel driver, adding printk statements to work out what was going wrong, and what values the probe routines returned. I worked out where in the code I could init the port, then got round to discovering what value(s) to write, and which port(s) to write them to.
I located the DOS driver that was loaded in Windows to initialise the card, without any documentation as to the switches to use. A quick inspection of the disassembly located the code that wrote from memory to the range of control I/O ports. I then located a table of IO Port value - bitmask pairs. Next to this table in memory was a simillar list of IRQ - bitmask pairs. By ANDing these pairs together I had the value to write to the IO register. I only needed the register itself.
I then searched the disassembly for a reference to the address of these tables - no such luck. Searching for addresses closer to this located the code that did the table lookup and wrote the correct value into memory. I hardcoded the write into my driver, reloaded the module, and all was now working.
I then spent a few minutes adding a couple of case statements and the odd debug message, then created a patch and submitted it. My patch appeared in the devel kernel a few weeks later.
Since then I've acquired a 6-port serial/synchronous card. Not only do no documents exist for this card, but there only code I can locate is a DOS diagnostic. The card was expected to run under a custom OS which I don't own. I'm not having much luck.
You are often tied to the specific config of the machine the driver was built on (usually the current RedHat). How many times have you run make config/menuconfig under 2.2.x, changed a single parameter, and discovered that half the kernel source needs to be recompiled?
The other problem with binary-only drivers is that the source has not been audited by one of the core kernel team (Linus, Alan or one of the others). You've no idea what the driver is doing. Is it susceptible to races? Does it tie up CPU time too much? Does it try to probe the hardware in a bad way? Does it use depreciated calls? Is there a memory leak somewhere? And so on. I've had enough of problems like this with add-on drivers for other Unix variants in the past to be very wary of 3rd party drivers.
My former company was into network support and software reselling, based in the UK. I once saw a presentation that listed customers. It included a couple of big companies, both of which had boughty the same piece of software from us in the past, 2-3 years ago, but no longer used it, as the manufacturers had abandoned all support and development.
They later sold a custom web server product without telling anyone that the backend was Lotus Domino.
The only problem I've seen with the current version (0.49 - released last month) was when resizing when connected to one particular Solaris 2.5 system over a slow and not too reliable connection. Occasionally the terminal session would still have the previous dimensions. Resizing when connected to an indentically configured machine, located on the same network segment as my PC, works fine.
I currently have it configured to use Andale Mono, one of the extra fonts bundled with IE5.0. The beauty with this font is that even at small point sizes all characters are distinct.
Oddly, IE5.0 on the Mac renders PNGs correctly.
Typical greedy bloody Micros~1, always adding new features to embrace and extend, without fixing things that are known to currently be broken.
I'd really like to remove all the IDE Bus-mastering code, and have the processor use PIO for everything.
Seriously, it seems that he doesn't understand the underlying flexibility of X - that the display and application can be in separate locations.
This is something that the Windows world lacks. So PC-Anywhere et al allow remote control, but these are limited to one simultaneous controlling user. (Ever tried to remotely connect to a NT box to run some admin utility that can only be controlled from the console, only to find that someone else is controlling some other application at the same time).
A window manager is a matter of choice - you don't even need one if only running one application under X. I currently run fvwm2, mainly to the small footprint and a cronic lack of disk space. I could be running sawmill/sawfish this weekend (he didn't mention that one), or maybe try the latest CVS snapshot of enlightenment.
His other main criticism was on the size of the code. The QNX installation he quoted is a very basic environment, I believe only using basic VESA modes with no acceleration. Yes it fits on a floppy, but then I ran Gem/DOS 3.2 from a single 360K floppy 11 years ago. X (and the Linux kernel itself) is certainly more complex in nature than the QNX demo software (which is only a demo after all, and not a fully featured OS).
Oh, and he got the name wrong - there's no such thing as X Windows. I have an instant distrust of journalists who get the name of software wrong - it never bodes well on the content of their articles.
Don't criticise what you can't understand!
I am certain they fiddled their reviews. Kit sent out for review was built on a different production line to a higher standard (eg cables were tied up, instead of left to hang loose). The casing was usually of a higher quality, and the parts generally top range. Paying customers (particularly the genral public), got whatever junk could be piled together in a day - often hardware that had been used elsewhere.
I am also sure I saw one of the 'technicians' activly running benchmarks for one machine in the test-lab, and these benchmarks were to be provided to one of the magazines. Oddly, this company usually scored very well in the benchmarks - I wonder why. (Personally, I regard all benchmarks with disdain).
I walked out of the place after 2 days. Not only was I disgusted by the way the company worked, but they treated staff like dirt.
The VINES os was based upon AT&Ts UNIX of the late 1980s, although with some modifications to the tasking mecahnism, and later the filesystem. Ports were made to SCO, AIX, HPUX, Solaris and other UNIX flavours (SINIX?), but due to a combination of bad marketing (always Banyan's weakness) and the fact that the UNIX port was based upon the features of the previous version of the native OS.
If configured properly, it worked well, and was used by many big organisations (Intel, Compaq, USMC). Some of these networks had around 1000 fileservers, all interconnected.
Banyan had a number of problems which contributed to their downfall. Hardware for their own proprietory servers (discontinued around 1991) was expensive. Marketing, as mentioned earlier, didn't exist. A port of their StreetTalk directory services was released when it clearly wasn't ready. Third parties were not able to develop device drivers until the mid 1990s (it took almost a year for a working server 3c5x9 driver to appear). Oh, and marketing from Redmond played a factor too.
One major contributing factor was the mail client. The DOS one worked fine, and was ported to OS/2 and Unix (the latter only working for shell accounts on a system also running as a Banyan server). Instead of developing their own simple one, Banyan bought out Beyond Mail. This mail system had rules processing, but these were never integrated into the process running on the mail server, and needed its own engine. The first Banyan labelled version of Beyond worked (sort of), but the next one was a mess. Slow, buggy, always crashing with 'assertion failed' errors, and not able to cope with the year 2000. There was one 3rd party mail client available that was still maintained - sites either switched to that or abandoned Banyan's email. Once Email was running on another system the integration of users as an object was lost.
The last straw was for Banyan to announce an agreement with Microsoft, basically stating that they would migrate all current users to NT. This too lost a lot of customers.
Banyan still just about exist, although only legacy customers remain. Most sales that I was aware of during the last 4 years were upgrades, with a few new servers to existing customers. The major players (Compaq, Intel etc) swiched over to NT (spit!)
If Banyan has invested some of their UNIX IP into offering a Linux port, they may still be going concern. They did announce an 'open' version of StreetTalk a few years ago, but I don't recall anyone actually getting hold of it. Some of the aspects of their protocols were very closed, particularly security, so a OSS clone would be difficult. Once OSS cloning was mentioned on the Banyan mailing list most customers had already given in the towel.
I became the only person in the UK supporting Banyan's VINES on SCO Unix, and had real problems getting support from anyone knowledgable through the official channels. I felt things could be better. and had to learn everything by trial and error.
In 1995 I was tasked with setting up our DNS server on an existing SCO system which ran other tasks. The BIND implemenation shipped with the current SCO UNIX TCP/IP implemenation was ancient even then, and just wouldn't work with our then ISPs DNS. A few days of attempting to get help from SCO got me nowhere, so I grabbed a PC destined for use as a print server, upped the RAM slightly, snarfed Slackware via FTP, and had a working DNS server within a couple of hours, which not only ran quicker, but also had all the bits and pieces that cost SCO user an arm and a leg (TCP/IP, NFS, C compiler, nroff and so on). I started to relalise the SCO was overpriced junk.
We were still an official reseller/integrator/whatever, and still receive the glossy magazines, although they didn't come to me :-(. 3-4 years ago I finally saw one on a sales-droid's desk. Despite SCO being the magazine's title, over 75% of the magazine content was NT - and this was an official publication. I knew SCO had troubles.
I'm just suprised that they've lasted so long. Only the actions of suits in preferring SCO over Linux or *BSD ('We need someone to blame when things go wrong! - Waddya mean no-one;s sued Microsoft') has kept them afloat for so long.
I want them to either go out of business, or to suffer the ultimate fate bestowed upon the dead - a buyout from Computer Associates.
Didn't the same companies team up 16 years ago with a certain software company to launch MSX?
Performance here in the UK sucks badly at the moment.
There was a report on a UK television programme a few years back concerning a diasabled woman who turned up to vote, but was told that her proxy had voted on her behalf. In the UK a proxy can be nominated to vote on your, but a voter can still vote before their nominated proxy, if needed.
It appeared that one of the local parties had been registering themselves as proxies for a number of local disabled people, then voting on their behalf. I don't think any charges were ever made, which was odd as the majority for this seat was very close.
It is the most unstable piece of commercial I have to use at work. Generally my NT Workstation has uptimes measured in weeks, mainly due to me ensuring that all device drivers are current. However once I run JBuilder I often have to reboot. JBuilder seems to want to write directly to screen memory during startup.
If I can get hold of the debugger copy of Windows NT I'll have a look to see what it does do. I did do simillar once with Notes under Win 3.11, and was astounded by the number of incorrect calls that POS was making.
The organisation I used to work for had a department that decided to sell a commercial Firewall (no longer in production, thank $DEITY). A copy was installed in our Office to see if it worked. The clueless fool who installed the thing didn't understand how FTP worked, and we were without FTP from a browser until my department got the firewall's password, when suddenly everything worked.
My department didn't have anything to do with the configuration at customer sites, but were called if things went wrong. The pilot site called a few weeks later to say that had been implicated as a source of spam. Of course, my team knew nothing about the config of this system, but were still expected to resolve it.
Digging around revealed that the SMTP process on this firewall was being used as an SMTP relay. We contacted the UK distributer of this product. We eventually got a replay, stating that the fix would be available in the next maintenance release of the firewall, currently scheduled to be released in 6 months time. Given that this was a major security hole, the fix should have been out as soon as the hole was discovered.
I don't recall the fix ever being released. I do recall around a year later a collegue stripping down all PCs in the building, looking for older versions of the 3C509 NIC to install in our filewall, as it was all that it still supported, so I'm wary if anything was fixed.
A while after that I received a call from a recruitment agency. The company responsible for the firewall had abandoned it, and now the UK distributers were now developing it themselves. They wanted someone to work in customer support. My response was two works, the second one being 'Off!'.
I've seen how lax many major organisations are when it comes to security (hey, we'll install a firewall, that will fix it), having worked with some sites that were a nightmare. Getting some customers to install ciritical patches was impossible until we refused to take any more telephone calls until they planned an upgrade.
Unfortunatly, there's not much scope in the security field at my current place of work (too much NT). Anyone with any ideas on how to progress - BTW I'm based in Yorkshire, England.
BTW the PC586 was a pain to configure, compared to the then current 3Com cards.
There was aslo something called the Hobbit drive, but I can't recall if this too used tape loops.
Are the data catridges used by Dec Vaxen loops? I saw one once, but didn't have the nerve to take it apart.
In 1982 Sinclair annouced Microdrives for the ZX Spectrum, although it took a year for them to arrive. The QL and ICL One-Per-Desk (a QL clone with integrated Lan and telephone) also used them.
They consisted of a loop of tape in a matchbox sized case. Capacity was approx 80-100K (wow!). Reformatting after use would increase the capacity due to tape stretch.
They were unreliable anf propriatory (only Sinclair made them, and they sold initially at 5 UKP per cartridge IIRC). They were also slow due to seek times in seconds.
Typical Sinclair - designing their own solution when cheap storage, such a 3.5" floppies, was already emerging.
I wonder if the name is still Trademarked??