The article was:http://thebetaguy.com/exclusives/?postid=1029344029 I am not sure about it's technical correctness.. AAP
On Mon, Feb 23, 2009 at 4:16 PM, mabshoff <mabsh...@googlemail.com> wrote: > > On Feb 23, 5:16 am, ahmet alper parker <aapar...@gmail.com> wrote: > > Hi, > > > I think one of the main idea behind developing a real alternative for > Magma, > > Matlab, Mathematica etc. is their license costs and restrictions on > altering > > the source code etc. Personally, if I had chance to not to use windows, I > > won't. Let me give an example, last night I read an article about > backward > > compatibility issues of windows 7 and I understand that they will not > > provide binary compatibility to previous windows versions, but they will > > supply an abstraction layer and api for previous versions which will run > > programs like a virtual machine like apple did for previous osX > > compatibility > > Do you mean the OS 9 compat layer in OSX? That is quite different than > the VMWare image in concept. > > > (if I understand it right). > > Which article was that? I am pretty sure you misunderstood something > since Microsoft is extremely interested in keeping old applications > working on new Windows releases. I.e. they bend over backwards to keep > code running that for example *writes* global registry entries or > opens global config files rw. All these things are rather obvious > mistakes not to make, but there is a lot of code out there written in > the assumption that you have admin rights or can write to various > system files. This is an obvious design pattern flaw in that user > code. The fact that the Windows ABI is stable is the greatest strength > and weakness simultaneously IMHO since it one hand keeps old binaries > running (modulo bugs :) and at the same time keeps cruft in the APIs > that people keep (ab)using. Windows did not only become dominant on > the desktop just because Microsoft played hard ball with hardware > vendors (Any other vendor in the same position would have abused their > monopoly in a similar way, i.e. all the Unix vendors got nearly killed > or badly hurt because they did the same tactics of abusing their > market position and tie their users into their system. Along come > Windows NT and ten years later had eaten nearly everyone's lunch > except that Linux also happened in the server space), but because it > had the software people wanted. Windows pre W2K and not based on the > NT core was *terrible*, but if you know what you are doing Windows is > a stable and securable operating system. The default config for say > Vista is still pretty insecure, but I recommend to install W2K8 and > use the build in IE in default mode. You will be surprised how locked > down it is. These days an amazing number of linux boxen get rooted as > can be seen if you look for example defacement statistics. There are > excellent and secure MS products, i.e. compare the security holes of > IIS and Apache for example. MSVC also creates significantly better > code than gcc on Windows, specifically the 64 bit target. This should > not be misunderstood that Windows is the better OS, in general > different tasks require different tools, but it would be idiotic of > use to ignore the 90+% of the desktop pie that is running nearly all > XP or higher. > > > My personal observation is that > > virtualization is going to be (of course they are not 100% the same > thing) > > the future. > > I doubt that. Some of the useful bits will be tightly integrated into > main stream OSes, i.e. I would not be surprised if in the future you > could start a XP instance in a post Winows 7 release. > > > So maybe, not a vmware version but a version built on a fast > > linux (with no necessary options of os) and a fast and lightweight > virtual > > machine with an msi installer to run it, can be better. > > This will still leave numerous issues, i.e the file system not being > integrated into the local filesystem transparently. There are various > Linux projects that provide linux as a process in Windows, but all of > them require admin access, which is a show stopper IMHO. You also end > up using way more space for example that way and Sage is fat enough as > it is. > > People in general are unwilling to switch from what they know and even > 5 years ago I thought I had to be out there convincing people to use > Linux on the desktop because it is better. I have long given up on > that desire and am perfectly happy to let people use what they want. > It is the best tool for the job, not some politically motivated > inferior solution in many cases. > > > Also, maybe not > > today, but when a new windows come to scene, again we need to alter our > > codes to recompile it to work with the new os, which is a waste of time > and > > effort (although ms is promising for win 7 for backward compatibility). > > That is not your call at all. I can work full time on Sage in part > because of the sponsorship of Microsoft Research. And you can ask > people around here what they would chose between me working on Sage > full time while also working on various ports or having to get a real > job and spending significantly less time on Sage. > > Also: Every time you port to a new OS or compiler you find plenty of > bugs and issues in the codebase. The Solaris port shook out a lot of > build bugs as well as issues in the code itself, so the port increased > the quality of Sage. The same will happen to a much larger extend > while porting to MSVC. I have ported major projects to MSVC from gcc > and I can tell you that the resulting code was cleaner and some > serious issues gcc did not detect were fixed. This kind of cleanup > happens regardless if you go MSVC to gcc or vice versa, but since > nearly all code in Sage as well as it components are developed using > gcc this is just the way it is going to be. > > > Anyway, > > these are just a user's opinion, so don't get too much angry on me :) > > I am not angry with anyone in this threat. If I had flame anyone in > this thread I am sure other people would have noticed ;) > > > AAP > > Cheers, > > Michael > > > --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
