If you were outside academia and research, no Unix was free in those days. BSD still had large bits of AT&T code and you needed their licence to run it which was far from cheap. It wasn't until FreeBSD that AT&T code was totally eliminated.
The other problem with BSD in the eighties was that it was fast but neither secure nor reliable. This even stretched to Digital's own Unix (BSD based), Ultrix-32 as it was then.
Digital's user group, DECUS did do a lot to help software distribution though and this was one of the springboards for the open source movement. Digital was also one of the original major contributor's to X development.
I was invited to a presentation on the VAX back in 78 (I think). I was already using PDP-11s and the system looked kind of neat and capable.
I didn't get to use the VAX until a couple of years later - but it was neat! We had no useful C compiler, but when the assembler (Macro-32) had higher level constructs than 'C' (even good exception handling in hardware), well what do we need it for. If you wrote some decent macros, you could even do structured programming in assembler.
In the earlier days, they shipped most of the VMS source listings on fiche. The code was good and the system worked. If the documentation wasn't good enough, you looked it up on the fiche.
The hardware was excellent with very very low rates of failure (other hardware producd by Digital sucked). VMS ran very well on it (BSD to a lesser extent, but it was more hardware independent).
Others wrote about VAX/VMS and VAXclusters. I won't repeat their claims, just to say the availability was incredible, particularly against the early Unix systems). All parameters were copied into system space and validated which made for good security and reliability before being worked with. Unfortunately, performance sucked with all this checking.
I still have a 3100/30 and a double headed 4000/60 which I occasionally play with. I even did the PGP port for VAX/VMS on the 3100.
TCO of an OS is cost of purchase less lack of availability. An additional factor is whether the O/S runs on a standard platform.
Linux runs ok on PC type hardware as well as more specialised architectures. This means the platform is as cheap or as expensive as the customer chooses.
If I have a commercial O/S and things go wrong big time, I have a problem unless I am big enough to apply pressure on the manufacturer. How quickly can your O/S vendor diagnose and correct an O/S level bug?
With open-source operating systems such as Linux, down time is reduced because if all else fails, there are many more people in the world who can debug Linux sources than those of IBM and Microsoft. The rest of us just don't have access to the source code. As far as undocumented changes are concerned, well, you can take care of an O/S in the same way as you do an application, change control!!! Binaries can be fingerprinted very easily, which you should be doing anyway on any mission critical system.
The last point is the O/S cost. I will be less general here. NT/Win2K runs ok but it prefers to have the functions split across several boxes. Yes, I might pay $5K for NT Server with bits (say BO Server), but I should also have a separate proxy server, I also should have a separate database server. Each time, I'm paying a lot for a server license, and both the initial outlay and the recurring costs increase. Remember with M/S OLP, you get a licence for two years only - then you pay all over again.
A solar farm is big - especially if you want to power a dozen or so PCs and keep them cool.
There are other micropower sources of varying capacity, but still a lot greener than a central power station. For example, how about small gas-turbine units driven by natural gas? They exist now and can give power from 25Kw upwards. With only one moving part (the rotor), they are very reliable.
There is a place for solar and wind farms, but you need a *lot* of space for them.
Many years ago, I worked in a building that was shared with HMV/EMI record publishing. They owned the labels through to the retail outlets.
I can confidently say that the $14.99 markup that was being made was being spent wisely on the cool dudes working there and launch parties. Since then, other overheads have emereged, such as paying music show producers for 'needle-time' and, shock-horror, the groups of kiddies who went into shops being sampled for the Top-50 statistics.
Also, you should remember the cost of all that coke (the powder, not the liquid)!
In St. Petersburg, Russia where illegal copies sell for $2, you can get real CDs for about $10-$15. If the record industry is profitable there, why not in the US and Western-Europe?
Sorry, I do run Win 2K, but before I get slammed, I definitely run Linux and OpenBSD.
The NT stack definitely used some of the BSD code as a starting point - you only have to look at the way some of the things work, but it has been heavily Windozed to function, and not all that well as the underlying architecture is continually evolving.
No - Win 2K is not OpenBSD and seems much too fat to run the kinds of load that a service like Hotmail offers. Then there is the matter of IIS 5 versus Apache, it too is bigger and apparently much slower.
As far as overall reliability is concerned, I'm not running any high volume service, just a general purpose development system, but Win 2K does seem more reliable now than NT4 was.
I just read through the arguments against Napster and want to focus on their 'allowable' copying suggestion...
Just say I had a load of copyrighted material, I don't have to use the original names, I can use what ever I want and then use a non-napster route (i.e. IRC) to inform the rest of the world that, for example, music by "Tenniebopper" was really another name for "Britany Spears".
Lawyers don't do technology and it shows. Neither do music publishing execs!!!!
Open source is good, but working with locals is even better.
I have worked on a number of projects where we pay locals to develop software under our direction that they need. They can often program well but certain technqiues may not be well known and project management definitely isn't.
The flip side of working this way is that you start to appreciate that a 486/66 with 32MB is still quite common in some places and different techniques are applicable which you can learn from the locals.
I spend about 50% of my time working for aid programmes in Russia and Uzbekistan. It does pay, but not particularly well, however it is kind of different and it sure beats a 9-5 environment or a cubicle.
Note, that I have always found it better to work for something (e.g. living expenses) than totally free. If your work costs the organisation something, then they will manage you better, even if in real-terms it is peanuts.
So select the part of the world that interests you then contact someone like the Peace Corps or even USAID. They may not be able to use you directly but they may have contacts with someone who can.
The Germans were able to read the codes used for communication between naval headquarters and convoys quite easily. This was because they started with a standard introductory text which provided a useful crib. Other codes were broken, but all Ultra intelligence was sent to the field using well-managed one-time pads.
Slashdot Mirror
If you were outside academia and research, no Unix was free in those days. BSD still had large bits of AT&T code and you needed their licence to run it which was far from cheap. It wasn't until FreeBSD that AT&T code was totally eliminated.
The other problem with BSD in the eighties was that it was fast but neither secure nor reliable. This even stretched to Digital's own Unix (BSD based), Ultrix-32 as it was then.
Digital's user group, DECUS did do a lot to help software distribution though and this was one of the springboards for the open source movement. Digital was also one of the original major contributor's to X development.
Nah, not a doorstop - more of a boat anchor. The older Vaxen had a lot of metal in them!
I was invited to a presentation on the VAX back in 78 (I think). I was already using PDP-11s and the system looked kind of neat and capable.
I didn't get to use the VAX until a couple of years later - but it was neat! We had no useful C compiler, but when the assembler (Macro-32) had higher level constructs than 'C' (even good exception handling in hardware), well what do we need it for. If you wrote some decent macros, you could even do structured programming in assembler.
In the earlier days, they shipped most of the VMS source listings on fiche. The code was good and the system worked. If the documentation wasn't good enough, you looked it up on the fiche.
The hardware was excellent with very very low rates of failure (other hardware producd by Digital sucked). VMS ran very well on it (BSD to a lesser extent, but it was more hardware independent).
Others wrote about VAX/VMS and VAXclusters. I won't repeat their claims, just to say the availability was incredible, particularly against the early Unix systems). All parameters were copied into system space and validated which made for good security and reliability before being worked with. Unfortunately, performance sucked with all this checking.
I still have a 3100/30 and a double headed 4000/60 which I occasionally play with. I even did the PGP port for VAX/VMS on the 3100.
TCO of an OS is cost of purchase less lack of availability. An additional factor is whether the O/S runs on a standard platform.
Linux runs ok on PC type hardware as well as more specialised architectures. This means the platform is as cheap or as expensive as the customer chooses.
If I have a commercial O/S and things go wrong big time, I have a problem unless I am big enough to apply pressure on the manufacturer. How quickly can your O/S vendor diagnose and correct an O/S level bug?
With open-source operating systems such as Linux, down time is reduced because if all else fails, there are many more people in the world who can debug Linux sources than those of IBM and Microsoft. The rest of us just don't have access to the source code. As far as undocumented changes are concerned, well, you can take care of an O/S in the same way as you do an application, change control!!! Binaries can be fingerprinted very easily, which you should be doing anyway on any mission critical system.
The last point is the O/S cost. I will be less general here. NT/Win2K runs ok but it prefers to have the functions split across several boxes. Yes, I might pay $5K for NT Server with bits (say BO Server), but I should also have a separate proxy server, I also should have a separate database server. Each time, I'm paying a lot for a server license, and both the initial outlay and the recurring costs increase. Remember with M/S OLP, you get a licence for two years only - then you pay all over again.
A solar farm is big - especially if you want to power a dozen or so PCs and keep them cool.
There are other micropower sources of varying capacity, but still a lot greener than a central power station. For example, how about small gas-turbine units driven by natural gas? They exist now and can give power from 25Kw upwards. With only one moving part (the rotor), they are very reliable.
There is a place for solar and wind farms, but you need a *lot* of space for them.
Many years ago, I worked in a building that was shared with HMV/EMI record publishing. They owned the labels through to the retail outlets.
I can confidently say that the $14.99 markup that was being made was being spent wisely on the cool dudes working there and launch parties. Since then, other overheads have emereged, such as paying music show producers for 'needle-time' and, shock-horror, the groups of kiddies who went into shops being sampled for the Top-50 statistics.
Also, you should remember the cost of all that coke (the powder, not the liquid)!
In St. Petersburg, Russia where illegal copies sell for $2, you can get real CDs for about $10-$15. If the record industry is profitable there, why not in the US and Western-Europe?
Sorry, I do run Win 2K, but before I get slammed, I definitely run Linux and OpenBSD.
The NT stack definitely used some of the BSD code as a starting point - you only have to look at the way some of the things work, but it has been heavily Windozed to function, and not all that well as the underlying architecture is continually evolving.
No - Win 2K is not OpenBSD and seems much too fat to run the kinds of load that a service like Hotmail offers. Then there is the matter of IIS 5 versus Apache, it too is bigger and apparently much slower.
As far as overall reliability is concerned, I'm not running any high volume service, just a general purpose development system, but Win 2K does seem more reliable now than NT4 was.
I just read through the arguments against Napster and want to focus on their 'allowable' copying suggestion...
Just say I had a load of copyrighted material, I don't have to use the original names, I can use what ever I want and then use a non-napster route (i.e. IRC) to inform the rest of the world that, for example, music by "Tenniebopper" was really another name for "Britany Spears".
Lawyers don't do technology and it shows. Neither do music publishing execs!!!!
Open source is good, but working with locals is even better.
I have worked on a number of projects where we pay locals to develop software under our direction that they need. They can often program well but certain technqiues may not be well known and project management definitely isn't.
The flip side of working this way is that you start to appreciate that a 486/66 with 32MB is still quite common in some places and different techniques are applicable which you can learn from the locals.
Hugh.
I spend about 50% of my time working for aid programmes in Russia and Uzbekistan. It does pay, but not particularly well, however it is kind of different and it sure beats a 9-5 environment or a cubicle.
Note, that I have always found it better to work for something (e.g. living expenses) than totally free. If your work costs the organisation something, then they will manage you better, even if in real-terms it is peanuts.
So select the part of the world that interests you then contact someone like the Peace Corps or even USAID. They may not be able to use you directly but they may have contacts with someone who can.
Hugh.
The Germans were able to read the codes used for communication between naval headquarters and convoys quite easily. This was because they started with a standard introductory text which provided a useful crib. Other codes were broken, but all Ultra intelligence was sent to the field using well-managed one-time pads.