Am Dienstag, 15. Januar 2013 22:47:06 UTC+1 schrieb jason: > > On 1/15/13 3:08 AM, Jan Groenewald wrote: > > Also if there is interest, a version can be made that installs Ubuntu, > > the source code for Sage, and all needed developer tools to build and > > run Sage. > > I'm really interested in a USB image that I can hand to students that > want to do development with me that: > > * I can modify (to include the Sage cell server, git, etc., for example) > > * have all the development tools installed, so that Sage can easily be > upgraded, rebuilt, etc. > > * can have further things installed by users as needed > > * can be installed to a hard disk if the student is interested > > Thanks, > > Jason >
I thought much about all this and worked long in that direction. Let me write a statement and bear with me that it might be a bit lengthy... All the things you ask for are possible with the current LiveCD version, and the possibilities were there since at least 3 years when I made an first image of sage version 431. At that time I had something in mind with very similar goals of Jan Groenewald and Nicolas M. Thiéry. I wanted to create a smart scientific Linux distro for easy distribution, be it as live CD, as USB image, or hard disk install. I had in mind especially teaching situations, where the whole system could be given out to students, and also especially situation in developing countries where it is expected to have lower average computing resources and just "old machines". Assuming all sciences need math it was logical to choose SAGE to be included into the base. All the features you ask for were there from the first release and the base being "Puppy Linux" was no accident. Reviews rate Puppy as "King of the Live distros" for small size, portability and hardware support from modern to very old computers. (e.g recent distrowatch review<http://distrowatch.com/weekly.php?issue=20121224#review>, "amazing" review<http://mylinuxexplore.blogspot.co.at/2012/10/puppy-linux-54-precise-review-amazing.html> ) Why I had the impression all the time of low interest of others, while it now seems that several people work with similar ideas on their own projects? I can just make guesses, but lets try it: * the base Puppy is different in several aspects and thats the reason why it is not perceived as serious enough as a scientific platform (see e.g. the comments of M. Thierye - first post here<https://groups.google.com/forum/?fromgroups=#!topic/sage-devel/PnqjVmpp6bU> ) * Maybe I failed in communicating my goals and by calling it "Live CD" nailed it and misguided people to miss the other potential (USB, hard disk install, developement, packages) of this piece of software. * I am not a member of any scientific community any more and work in a different field, so this was just a completely private and effort. I missed especially the possibility of direct feedback, opinions and testing by others. * I had time for this projects during some months in fall and winter, so there were gaps of more than half year with no development and reduced communication to interested people (I also lacked proper internet connection during that time) * During the last 2 years I concentrated on several aspects which seemed logical for me with the small base I had, especially on building virtual machine images and lobby for it here on the google.groups since I felt it is the only realistic solution to the "sage on windows" dilemma. In going this way I neglected the original goals of my project. * In my discussions I crossed opinions with some of the most respected sage developers (Robert Bradshaw, Volker Brown) on the matter of size. While it is true that size of the software doesn't matters in a 1st world environment with up to date powerful machines and unrestricted, super fast internet access, size is a matter on older computers and in an environment with shaky internet access, low download speeds and old, RAM and disk challenged machines. Right at the moment I see 4 projects maintained from this community with similar goals or structure: 1) Nicolas M. T's distro based on debian live 2) Jan Groenwalds distro based on Ubuntu 3) my SAGE Live CD distro based on Puppy Linux 4) the Fedora based image of Volker Braun to provide the base for the Sage virtual machine, this is maybe not geared toward a Live CD/or USB solution but still it is a complete custom linux distro including sage. Additionally there are other projects with very similar goals. E.g. an old but now dead project was the Knoppix based "Quantian" Live distro which still ranks high in the google list if you search for "Linux Mathematic". A recent project is the MathLibre Projec <http://knoppix-math.org/>t which also produces a Knoppix based live DVD (INCLUDING SAGE!) - last release was 2011. In a more broader sense (and because I am technically interested in it) there are also special Linux distros for simulation and numerical computing (e.g. CAELinux <http://www.caelinux.com/CMS/>). First obvious point I want to make: There is a lot work involved with such a project. Maintaining it, building a self supporting community which shares the workload, writing documentation, providing support and help, communicating about and advertising it. If it is gonna be successful than it will need more than one man teams. Once you release the work tends to increase. And very often you are confronted with annoying little things like missing keymaps, language support, printer or video drivers, wifi connection etc... Second point is: It is difficult to meet the broad range of expectations and application cases, there will be problems ranging from failing hardware drivers, language support, keyboard support, installation problems etc. As an example let me quote a response I got from a Guy from Iran. rasul wrote: > Hi emil and thanks for your great job, I have used your previous sagelive > and now I am managing to use your new one for my dissertation, 50 MB usage > of RAM is very fantastic! However, I have also a question. I'm going to run > my code on a quad-core machine with 18 GB of RAM, fortunately SageLive > supports multicore but does it support also 18 GB of RAM? If not then do > you have any suggestion? My code is so heavy that it failed on 4GB Mac > Machine, now I want to use my professor's machine and I want to make sure > it will work before to tell him I want to work on his machine. > > Thanks again. > I had to tell rasul that my Live CD (32 bit) was optimized for low specs and for the 18 GB of his professors machine he would need a 64 bit distro (forget about PAE). So my conclusion is, you can't meet all expectations, at least not with one version. >From my Point of view the approaches of Jan Groenwald with the Ubuntu base, Nicolas approach with a variant of debian live are very similar approaches. They use well established distros with good repos and an established user base. While Ubuntu has the polish, debian has the flair of beeing not commercial and really free. Volker Brauns Fedora machine is a similar approach and in a similar class of distro, but at the moment lacking the component of making a Live USB or CD from it. I don't know how much work it would be to add such a facility, but I guess it is possible in principle. My Sage Live project is different in the aspect that it has an unconventional base and has some rough edges. The biggest disadvantage is that it might be harder to get into it than into the established and polished big distros and that it might be more difficult to use the repos and add software. The advantage is that is unbeatable? small and it is really optimized to be a live OS. It is logical that everyone will choose a distro base which he/she feels most comfortable or which will seem "the right one" depending on some more or less objective criteria. I estimate it unrealistic that the people involved will completely abandon their "baby" to join another project, but in the long run it would be better to have one project together than so much effort split. For example last year I also had the idea to change the Sage live CD base from puppy to debian live. I had a 220 MB base of debian live (with XORG, JWM window manager and without sage) which made it realistic to add sage and keep under the 700 MB CD limit. It had some real advantages (working apt-get, real multiuser support) but I abandoned it at that time, because the puppy base was so much more polished and functional that it just didn't seem worth the effort to go in that direction. So may final suggestions are: *make the creation of a Scientific/mathematical USB/DVD distro as specified in this thread by Nicolas M. T and Jan Groenewald an official goal of the sage community (or part of it). *Have a look what the MathLibre / Knoppix people are doing in their project *On the sagemath homepage, change the download section to be " Live CD / USB / DVD images" *Offer different images and add explanations and documentation what the /specs/ advantages/disadvantages of those different images are. *Keep in contact, keep discussing and testing and maybe people can converge towards on base distro in the future. Cheers emil -- You received this message because you are subscribed to the Google Groups "sage-devel" group. 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. Visit this group at http://groups.google.com/group/sage-devel?hl=en.
