You are correct, it does use NT kernel, not just raw HAL. I really meant that it was at the raw NT kernel (file system, memory management and process scheduler only) and not built using Win32. It really is a Unix API instead of Win32 API for NT. As a matter of fact, Steve Walli said that they couldn't implement true Unix file hard links but only soft links because of using the NT file system.
The Ottawa Unix Users Group had a presentation last year by Stephen Walli, the chief architect of OpenNT (the original name of Interix product). He indicated that it worked by basically replacing the WIN32 subsystems of NT by separate code that was built on the NT HAL (Hardware Abstraction Layer) not Win32. This allowed Interix to bypass the inefficiencies of standard NT but run Unix software on same machine at same time as Win32 code. If Microsoft is buying this technology, it indicates either they want to bury it or that they need the Interix efficiency to gain in the server market. Since one of the chief complaints about NT as a server is its Win32 based complexity and lack of efficiency, I would bet on the latter. Trying to bury it would just open the market up for someone else to make software that runs Unix or Linux under NT. It is much easier to port *NIX to NT than the reverse because Unices are better defined and better partitioned. As the WINE project shows, they are lots of hidden parts of the Win32 API, and even MS probably doesn't know all the gotchas.
I have been programming for 35 years and have taught several programming courses. One of the best books I ever came across for learning programming was George Polya's "How to Solve It", originally written in the 1940's. It didn't have much about coding (since computers weren't around when he wrote it) but it did have a lot about analysis of problems, frameworks, abstraction and most other tools of modern programming. In one sense, introductory programming should not need computers at all. Computer programming has two main tasks. The first is developing algorithmns, the second is developing data representations. Each task affects the other but both are neccessary for correct implementations of systems. Developing algorithms for hand implementation (playing computer) helps you understand things much better than trusting any language implementation. Drawing diagrams of the relations between data elements is neccesary before declaring any types and objects in a language.
Only once someone understands iteration, alternation and object properties on paper should the computer become involved. Thus the language should be as close to formalizing natural language descriptions of problems as possible.
Python is a good language for algorithmic analysis because of a clean clear syntax, although dependence of indenting for blocking can sometimes confuse people. It is not such a good language for data type analysis because it allows mistakes in types to stand in programs. Python with a stronger type system would be a good teaching language, especially with the case insensitivity of Alice.
As a sysadmin, my general problem with downloading porn is the volume and time. When I find 2% of my Internet bandwidth going to a porn site, I don't really care about the content but the waste of bandwidth going into someone wasting company time and money. I think too many people see Internet access at work as a free ISP rather than a tool like the telephone.
I don't think anyone would see anything wrong for a company disciplining somebody who had a huge long distance bill to 1-900 porn sites. Why should they not halt time, space and bandwidth waste of their Internet access?
I monitor Internet use looking for bottlenecks and bandwith problems. I would rather save bandwidth by stopping access to porn sites than access to system upgrades.
Slashdot Mirror
You are correct, it does use NT kernel, not just raw HAL. I really meant that it was at the raw NT kernel (file system, memory management and process scheduler only) and not built using Win32. It really is a Unix API instead of Win32 API for NT. As a matter of fact, Steve Walli said that they couldn't implement true Unix file hard links but only soft links because of using the NT file system.
The Ottawa Unix Users Group had a presentation last year by Stephen Walli, the chief architect of OpenNT (the original name of Interix product). He indicated that it worked by basically replacing the WIN32 subsystems of NT by separate code that was built on the NT HAL (Hardware Abstraction Layer) not Win32. This allowed Interix to bypass the inefficiencies of standard NT but run Unix software on same machine at same time as Win32 code.
If Microsoft is buying this technology, it indicates either they want to bury it or that they need the Interix efficiency to gain in the server market.
Since one of the chief complaints about NT as a server is its Win32 based complexity and lack of efficiency, I would bet on the latter. Trying to bury it would just open the market up for someone else to make software that runs Unix or Linux under NT. It is much easier to port *NIX to NT than the reverse because Unices are better defined and better partitioned. As the WINE project shows, they are lots of hidden parts of the Win32 API, and even MS probably doesn't know all the gotchas.
I have been programming for 35 years and have taught several programming courses. One of the best books I ever came across for learning programming was George Polya's "How to Solve It", originally written in the 1940's.
It didn't have much about coding (since computers weren't around when he wrote it) but it did have a lot about analysis of problems, frameworks, abstraction and most other tools of modern programming.
In one sense, introductory programming should not need computers at all. Computer programming has two main tasks. The first is developing algorithmns, the second is developing data representations. Each task affects the other but both are neccessary for correct implementations of systems. Developing algorithms for hand implementation (playing computer) helps you understand things much better than trusting any language implementation.
Drawing diagrams of the relations between data elements is neccesary before declaring any types and objects in a language.
Only once someone understands iteration, alternation and object properties on paper should the computer become involved. Thus the language should be as close to formalizing natural language descriptions of problems as possible.
Python is a good language for algorithmic analysis because of a clean clear syntax, although dependence of indenting for blocking can sometimes confuse people.
It is not such a good language for data type analysis because it allows mistakes in types to stand in programs. Python with a stronger type system would be a good teaching language, especially with the case insensitivity of Alice.
I don't think anyone would see anything wrong for a company disciplining somebody who had a huge long distance bill to 1-900 porn sites. Why should they not halt time, space and bandwidth waste of their Internet access?
I monitor Internet use looking for bottlenecks and bandwith problems. I would rather save bandwidth by stopping access to porn sites than access to system upgrades.