On Sun 28 Sep 2003 at 08:19, Warly wrote: | It may be a good idea, before cooker opens again, to take these | days to have some brainstorm. | | May you give your opinion on : | | - What was wrong in 9.2 development process?
1. N America mirrors... out-of-date, inaccessible with too many already logged in (priority to on-campus users). No local, reliable sources of snapshot CDs (I'm stuck with dial-up Internet access so I need a new snapshot on CD when there are major changes, then I can keep up for a while with overnight sync'ing until there's another major--as in size-- change). 2. Too little time. Start earlier. I keep older and unusual boxes for testing: PI system, GigaPro MB with 1Gigapro CPU, slot-A Athlon on Amptron MB. Home network built from usable parts of 'throw-away' PCs, 'final closeout' bargains and lease return bargains. With enough time, I can try CD and network installs on each and perhaps find and/or fix bugs. | - We though a bit late in the 9.2 developement process to split | cooker ml, we should do it now. IMHO, do _NOT_ spilt the mailing list. Strongly encourage subject tags for sorting with email client filters. [cooker] to a directory, [Bug xxxx] to bugs sub-directory, etc. Put out a guideline message once a month or when there's a change. Find a volunteer to create and maintain a hierarchical list of recommended subject line tags to sort on. I like being able to scan all the traffic. Most gets deleted after a quick look. I won't own an IA64 or Ath64 anytime soon but I want to know what's going on with the whole distro and the company. I also like being able to use filters to sort the mail that interests me most into a directory that I will look at first. For each release cycle, strongly encourage tags after [cooker] to clearly indicate to which snapshot the message applies: [beta1] -> [rc3]. Follow with the common name of the tool or package: [install] or [kde] or {apache2] or [whatever]. Use filters to sort things out as you prefer. | - What could we do to improve 9.3/10.0 development. 1. START EARLIER. 2. Set clear goals for 9.3/10. Not just dates but priorities and objectives. For example: A. Install and initial boot flawlessly on the following: 1) off-the-shelf PCs {list}; 2) laptops {list}; 97% of other i586 off-the-shelf PCs; and over 90% of home-built and custom-built i586 PCs. B. Support the following SATA-enabled main boards {list} and drives {list} and 90+% of other available serial HDs. C. Include the latest stable version of {list including OO, KDE, Gnome, Apache, etc.}.... 3. I largely agree with James Sparenberg <[EMAIL PROTECTED]>, especially a focus on tools and a separate 'test' area on the mirrors. Emphasize the installer. Nothing frustrates a new (or even an experienced) user more than the installation repeatedly freezing or getting nothing but 'kernel panic' when trying that first boot. Too many give up in frustration and either go back to Microsoft Windows or try a different distro' (and then come back to message boards and post about Mandrake problems and how SuSe or RedHat or whatever worked fine). 4. Involve MandrakeClub members again and more. Months before the release cycle gets moving, create a place to vote for 9.3/10 features and packages in the Mandrake voting site. Start with a table showing all of the core packages. Tell them that KOffice, OO, Apache, etc. are part of the core and that the latest stable version of each of them will be in the release. Don't propose them and if someone else does, don't waste your vote for them. Clean up the distro, moving some traditional packages to contrib even though some will yell and scream and cry about the editor they used back in '92 being removed. Post a table for members to vote for which of the 'very nice' packages they prefer. The proftpd versus pureftpd thread is an example. Then allow the chaotic free-for-all of requests for packages and features. Let MandrakeClub members know that they can request a 'feature' as long as they can describe what they want. | - What should we do to improve the Wiki. If it ain't broke, don't fix it. It may break as it fills but it's OK for now. When cracks start to show, fix it or rebuild it or reorganize it or replace it or whatever it takes to make it work well again. | - Should we have cooker snapshot ISOs? Maybe short (two day?) freezes between cycles asking folks to try a fresh install, some building bootable CDs and others doing a network install, to check a change in installer or tools or a new default kernel or ??? | - What could we do, as a community, to increase the acceptance | of mandrakelinux? 1. First and most important, keep producing a Linux distro that's easy to install, configure, and update while retaining the stability and security of Linux. Continue making administering networks easy, increasing productivity of those who use it. Give CDs away to everybody who shows an interest. This msg is alrady too long but... climbing onto soap box... My college sophomore daughter started using my Mandrake desktop in high school. She got her boyfriend to try it. He was soon a Mandrake user. Another boyfriend taking a computer security class told me how his professor taught that Microsoft is inherently insecure and that Linux or BSD were needed for secure sites. The professor recommended Red Hat. I gave that boyfriend a set of Mandrake 9.1 download CDs to try. Now the whole class is using Mandrake because it's easier to install and set up... and the professor converted easily because it's rpm-based. Now the University of Deleware ftp server hosts an unauthorized but well-used directory of the MandrakeClub rpms. See ftp://ftp.lug.udel.edu/pub/linux/distributions/mandrake/devel/unsupported/ Unfortunately, their /cooker/ mirror was last updated in August. -grumble- MandrakeLinux won't see much difference this year. I'll buy a couple of 9.2 boxed sets as Christmas gifts making no real difference for Mandrake. I suspect that copies of the CDs will spread through universities in Corpus Christi TX (my son's area) and Newark DE (daughter's). Students don't buy boxed sets anyway and groups of students probably share one MandrakeClub member- ship. But note that Mandrake, Slack' and Debian are now the only distro's being mirrored by the UDelLUG. No RedHat!?! Fast-forward three years. Mandrake is better than ever. Those students are working in entry-level IT or junior management positions, frustrated because Microsoft or even RedHat or Debian are used by the company. They remember the stability, the easy setup and administration, graphic tools that make them more productive, and everything else Mandrake. One or two will sell Mandrake to their bosses and Mandrake will profit a bit. Now fast-forward a few more years. Those students have families at home and more responsibility and authority at work. They KNOW about total cost of ownership, ease-of-use and productivity enhancement from first-hand experience. I pity Microsoft vendors in five or ten years. Today's students will be buying DVDs or boxed sets because they don't have time to do downloads, check dm5sums and such. They can afford the convenience of the DVD for thier home PCs. More important, their companies can afford Mandrake enterprise products and those students will be in management positions. They KNOW, from first-hand experience, that Mandrake will save their company money and make their IT more productive. Buy your shares now while they're affordable, then cash in and retire when Mandrake rockets upward around 2010. So, most important, is continuing to make Mandrake a super product that's stable, useful, easy to install and configure, and everything that makes Mandrake Mandrake. ...stepping down off soap box after singing to the choir... 2. Name/Logo recognition is needed. Boxes on shelves will help. Mentions on other web sites, too. Advertising when Mandrake can afford it. Individual users carrying download CDs to install fests and helping those switching from MS Win get started will make a little bit of difference every time. Put a Mandrake logo from http://www.linux-mandrake.com/en/poweredby.php3 on personal web sites. Get case badges and display them. The only current source I know of: http://www.cheeplinux.com/cat.pl?session_id=7 | - How to have more contributors? 1. Define and list what contributions are needed with a list of necessary or recommended skills. Personally, my skills are limited and how much time I can devote varies seasonally. I don't know if I can contribute anything of value other than occasionally confirming a bug. Documentation, maybe? 2. Sponsor/host on-line classes? A mentorship or apprentice program? Expand and update MandrakeClub 'Campus'? More and better documentation and guidelines? For now, I want to learn DocBook and how to build Mandrake rpms properly. I found lots of documentation but much of it seems outdated. Some is overly simple and some assumes that I understand far more than I currently do. I want to package Dr Berstein's ClockSpeed and TaiClock ( http://cr.yp.to/time.html ) complete with post-installation scripts or maybe a Python wrapper/GUI tool to ease the set-up and a detailed HOW-TO. In my self-study, self-teaching mode, I'm sure to make dumb mistakes that might be avoided if I took some classes and had a mentor to review my efforts before I upload a trial/test version. At the rate I'm learning, I may get them packaged before the 2036 when NTP's counter overflows or 2038 when Unix time overflows. Will 2.6 kernels include TAI64 capabilities without a custom compile? I generally understand XGML and XML, how to mark up a document. In the past, I wrote lots of HTML with text editors. But how do I actually use all of those docbook style sheets that I installed? I can write. I can tag. I just don't know how to put it all together, from a practical standpoint. Do I just open Vi in a console and begin typing a declaration? But which one? Why? 'DocBook: The Definitive Guide' isn't as clear as Mandrake's DG. | And anything related to the mandrakelinux distro. [OFF TOPIC] Managers fear Linux, especially Mandrake Linux, because of experiences with Microsoft. Each version of Windows is, in critical ways, incompatible with previous versions. The forced transitions are disruptive and expensive but occur only every 2 or 3 years. Mandrake versions change twice a year, a terrifying prospect for an executive who was just forced to upgrade his Microsoft 'seats' to Win2K Professional and buy and install and configure all those new applications like Office2K and Norton Anti-Virus to protect it and... [BACK ON TOPIC] Split the 'Mandrake 10 ProSuite Edition' from the basic. Show commercial desktop/office users that Mandrake won't stop supporting them and won't force them to upgrade.