OT: Antivirus and other antimalware stuff

[Andy Burns]

Yes and still growing.

Nov2010-Jan2011

Apr2011-Jun2011

[Andy Burns]

?

[John Rumm]

More to the point, make sure you have standard terms and conditions that discount any liability for data loss.

[John Rumm]

Sorry, did not read all of your post... ;-)

(great minds think alike huh!)

[Huge]

Good point.

[Huge]

According to some reasonably trustworth figures I saw last week, they have about a 1/3rd each. Everyone else shares the remaining third.

[Nightjar]

I go back to when it was about the only way to own a computer, unless you had the money for a Commodore PET. In more recent times, it has been because off the shelf computers simply didn't do what I wanted. For example, before the need to put multimedia PCs in the living room created a market for virtually silent cases, the fans from five PCs, a server and a network hard drive made a fairly small office sound like the apron at Heathrow airport. So, when the time came to replace them (more power, more fans) the next lot went into a 19" rack in the warehouse, with enough battery backup to run everything for about 15 minutes, and they were connected to the office with KVM cables. Result: an office you could hear a whisper in.

Colin Bignell

[Mark]

HP-UX used to crash back in the early 90's.

Windows memory management is somewhat odd in particular cases and I have been asked to solve issues relating to this on more than one occasion in the past.

[Bob Eager]

With one or two exceptions on the memory management front, notably Mini- UNIX (TM).

I'm not seriously arguing, but I thought it was vaguely interesting)

I have a copy somewhere, and I installed it once.

[Huge]

Not "allways". Unix ran on many PDP-11s without memory management hardware.

[Man at B&Q]

Which version of XP? 32 bit is limited to 2G per process.

MBQ

[John Rumm]

I think even then it was still "resource managed" even if there is no hardware support to enforce memory protection. So if an app requests a lump of system memory and then quite without freeing it first, unix will come and mop up after it and free the RAM rather than suffering a memory leak.

[The Natural Philosopher]

Precisely so.

A friend who used to write that stuff managed to do some kernel meomory allocation/deallocation in NT, quite by accident.

He noticed the free pool went down by IIRC 16 bytes every time.

He wrote a daemon that would crash any NT kernel in a few hours using perfectly legal code.

The same code on Linux/Unix simply made no difference to anything.

[The Natural Philosopher]

Ah. That is probably the issue then.

[David WE Roberts]

Oh, and have you heard the urban legend that the reason for the name NT is that the letters WMT are one letter later than VMS and that a lot of NT development was done by ex-VMS developers?

Should be good for a geeky pub quiz :-)

[Bob Eager]

Just a development of the urban legend about the computer in 2001: A Space Odyssey. It was HAL; take the 'one letter later' in each case again...

[Huge]

Don't all O/Ses work this way? (I have no knowledge of the insides of Windows.)

[Huge]

Kinda. A lot of NT development was done by Dave Cutler, who worked on VMS and wrote RSX11.

[John Rumm]

Not all no, although it is becoming less common for them to not have it these days. Old versions of windows (i.e. anything prior to NT) could leak all over the place and get memory fragmentation problems, and other resource leaks. Some real time (and other very fast OSes) would eschew resource management and or garbage collection in favour of rapid response and low interrupt latencies.

[Bob Eager]

Well, not on Mini-UNIX. It only had one process in memory at a time, and that's all the memory it got. Programs were all linked to the same fixed base, and loaded 'over' each other. Of course, 'malloc' would work within the process, but I don't think 'malloc' was 'invented' then.