Jump to content

just an idea, seeing if any others may be interested...


tyme
 Share

Recommended Posts

It doesn't matter if people say no becuase this will be shi* loads of work and you, me, them, us will be posting all sorts of questions on here....

 

Big question:

Is this a leaning experience or serious attempt to make a distro....

In all seriousness I think the formwer is more realistic.

 

No one distro will do what all the developers want form 'their' distro so if we look at it like this then we are principally learning.

 

It needs to be phased. Like mini-distro first we can all agree on and then expanded later. Also an aim, like movix will keep us all on track..... it needs a mission statement type thing...

Link to comment
Share on other sites

  • Replies 64
  • Created
  • Last Reply

Top Posters In This Topic

ok, i have a bit of time i can squeeze out here at work and give you all an idea of what i was looking at, taking all your comments into consideration:

  • Fluxbox default WM-or another *box

 

[*]It's own packaging system. I don't like RPM. However, apt-get is intriguing.

 

[*]RXvt/Eterm are good, I like them both.

 

[*]Mozilla Firebird (is this the one you meant static?)

 

[*]Gaim (takes care of a more than enough IM services in one)

 

[*]Eroaster (or k3b....)

 

[*]Xine and XMMS (i like the standards) + Ogle for DVD (my personal fav)

 

[*]OOo, of course....

 

[*]Ethereal, nmap, and some wireless tools

 

Those are my basic ideas for the contents. I'm flexible in area-of course. However, in my opinion, KDE based apps aren't good to be used outside of the KDE environment. I don't like having DCOPS running for no good reason....maybe that's just me :-). BUT, what could be done is we have a standard setup on the CD and then if one (or a group) of us prefers a different app for a certain function we could make a package of that app and place it up on the server. I like the livecd idea, especially statics idea of creating on which contains the person preferences. That would be a cool tool.

 

I think the best plan for the basic install CD would be to try to keep away from making the GNOME/KDE choose for the user. This may be painful however...we'd have to see how it works out.

 

The packaging system: I was thinking pre-built packages optimized for i686 (like arch does). Possibly a mirror of those packages optimized for i586, if there is enough demand-and we have the support. Also, it would be a CLI GUI type of deal (especially for the install process-possibly a regular GUI later). Use of ncurses. Packages divided into groups-or the ability to choose different views. Not sure...I think grouped would be best. Searchable, of course. Those interested are welcome to add to this or come up with other suggestions. I'm not a big fan of RPM, as I said. apt/apt-get is interesting, though I haven't used it all that much.

 

One issue on my mind is server....I don't have the neccessary $$ to purchase a server for use...atleast I don't think I do. I'm a college student and all-funds can be tight sometimes :wink:. Not sure how we could resolve this issue.

 

Other thoughts: I'm going on the LFS adventure this week/weekend. I'd suggest the same for anyone really interested in this who wants to be a part of the core development team. I'm sure by monday my mind will be fried. I may need next week to recover :wink:

 

sarah31: I'll consider taking some time to talk to apeiro and/or speaking with other developers. This will be after my LFS experience gets done...I want to know enough that I won't get lost in convo :-)

 

everyone else: glad people are interested....hmmm...dunno what else to write :-)

Link to comment
Share on other sites

Big question:

Is this a leaning experience or serious attempt to make a distro....

In all seriousness I think the formwer is more realistic.  

personally, i'd like to think it's both. what's the point in making a distro if you don't intend for others to use it? it will, of course, be a learning experience at first-but-i hope eventually it will develop into even more.
Link to comment
Share on other sites

a few more thoughts that just popped into my head: there would be a nice GUI tool for configuration, _but_ I'm thinking it would just give you easy access to viewing and editing the config files. these files would be well commented so users could understand the idea. maybe an option for a simple config could be present...in case they don't want to learn every last tidbit....WDYT?

 

addition: one good thing is this would give those involved a better understanding of linux and therefore we would be better able to help people out on this forum :-)

Link to comment
Share on other sites

a few more thoughts that just popped into my head:  there would be a nice GUI tool for configuration, _but_ I'm thinking it would just give you easy access to viewing and editing the config files.  these files would be well commented so users could understand the idea.  maybe an option for a simple config could be present...in case they don't want to learn every last tidbit....WDYT?

 

Maybe something liek shorewall-config for ip-tables, but then easier....??

Suppose that all servers should eb configured safe from the start......

Link to comment
Share on other sites

Guest c_m_f

At first I was interested, not in a core dev kinda way, but help, i cant program, i don't know how linux works too well, but there are other jobs, but at the moment it just doesn't seem me, but thats because i love GUI (you said about GUI/CLI)

Also im a very definate KDE user, have tried Fluxbox, but like icons on my desktop that work, like ease of use, and like KDE in general, thats just my opinion though.

 

I think the live CD idea is great though, also something i read about was the idea of having important tasks which link to none GPL stuff (like Nvidia drivers, DVD stuff) is a simple link in the taskbar saying something like 'Install Nvidia Drivers' (if an Nvidia card/mobo is present) and then fetches form distro's package archive.

 

On the package front i've been losely following the development of AutoPackage which is a new package system. take a look at that, could even help in that project to help it make it off the ground faster as packaging will be prioity here (needs to be installed, and embracing this system early could play a large advantage given it is revsing what is wrong with linux packaging at the moment.)

 

I like the idea, espically i686 optmizitation (and i586 mirrors), it would be nice if on an install, you get install from cd options, then if no suitable pacakges are there, you can connect to net (if possible before packages are installed?) and d/l the ones you want eg. KDE, QT/GTK versions of Config programs, Gnome, etc.

 

I think the Config wizards/programs are a must, the one thing keeping me away from Gentoo is Config, i dont want to have to touch config files for a first off the bat install (fair enough, if i want to delve in to change some obscure value, but not fully configure X or something similair).

A great idea would be a good wizard for Samba, as i still havnt gotten it running and SWAT seems to cryptic, i just don't know what to do.

 

Well those are my opinions, obviously from a more newbie beneficial (although i dont calls my self a newbie, i have sympathy, and o distro at the moment aters for them) point of view (and KDE loving) but still i hope they are appreciated and taken into conideration, because to be honest i don't want to be involved in a project whose's end result i won't use.

 

Edit:

 

Fixed many typo's.

 

Also had another thought, how about having multiple versions, like 'home' or 'begineer' (features Kde/Gnome which ever is chose best for begineer) which also features an option to 'upgrade' the install to another catagory, eg Development, Professional, Power User, Intermediate, Enterprise, Server, Office, Multimedia/Gaming, Custom or a mixture of all of these (there are obviously too many options there, bu tcould be streamlined to less, as maintaining al of these would be horrible!).

These will contain packges appropriate to each type of install, and taylor the OS to the user more, and also still cut down on multiple packages.

 

Also When installing if oyu pick multiple Progs for same use, you have an 'evaluation' period, which you try all, and at the end pick a favourite and the rest are removed, tayloring the install to the user (havn't thought about multiple users yet though).

Link to comment
Share on other sites

OK tyme - did you read my post?? <:D It sounds like you have the exact same ideas as me but didn't bother reading my post ! LOL j/k

 

This is hilarious. Two weeks ago I read a shitload on LFS, a while before that I was trying out keeper linux which taught me a lot about live cd's and el turrito. Sounds like we were meant to be tyme. (Oh, and don't let sarah31 discourage you - it'll take us a while, but we'd have a lot of appreciated help) I know a lot of C and C++, as well as kylix for nice linux GUI apps like a config center.

 

I have to say, though, that coming up with our own packaging tool is not only a shot in the foot because everyone be limited to what we added to the package tree, but that leaves LSB in the dust. Like sarah, I'm all for standards, and like sarah, don't know about some of the LSB's decisions (especially rpm >:| ) but it would be really nice to have that opengroup symbol on the cover :) I'd like to support the LSB idea to encourage gaming companies, hardware manufactures, etc...

LSB Certification is for products meeting the LSB and OpenI18N specifications.

PLUS it'd take way too long to design and implement a new pkgtool, troubleshoot / bugfix it, then maintain thousands of packages ON TOP of designing and maintaining the distro. Let's start with the distro :)

 

RPM isn't so bad with urpmi. If fact, I kinda like it when I've got urpmi at my disposal. It'd be easy to have both our own urpmi-type app AND apt-get in there, which is easy to do.

 

OK on to some basics. Looks like we agree on fluxbox, rxvt/eterm, (is it?) firebird, Gaim, eroaster and CD Bake Oven, ogle/xmms/xine, OOo, {ethereal, etherape, nmap, snort w/swatch & syslog-ng, ipkungfu & Guarddog + Guiddog, you know, network basics}. I really want krusader (or equivalent) as the default file manager. It is by far the fastest way to manipulte files. I take it considering we are both college students this is slightly aimed towards college students, so I suggest we include any libraries that gamers would need to be able to install games later. Once we get the distro-ball rolling, I say we throw in some amazing free games (for example racer looks AWESOME and it's only like 10 megs I think).

 

I agree with a lot of what people said here, except the XBox thing. I don't own one and never will.

 

Something I am going to insist on is a lot of good documentation (quit groaning ;) and I need to learn how CVS works...

 

More fundamental, what is this distro aimed at? Speed/Gaming or stability/server. Gentoo-like install from source or pre-compiled binaries??  

If its binaries RPM or APT...? portage-alike?

This is a good point. I'd say it's aimed for college students, not quite so much a stand-alone, X-less server (other distros already do that so well). I wonder if we could write scripts for the install that compile from source and all the source is on the install cd... aru is really awesome with bash scripts, too. He might be able to give us a suggestion here and there.

 

Anyway, if we were to plan from now til X-mas time, I'll have a server starting right after x-mas ;)

Link to comment
Share on other sites

Guest c_m_f
.

 

I have to say, though, that coming up with our own packaging tool is not only a shot in the foot because everyone be limited to what we added to the package treePLUS it'd take way too long to design and implement a new pkgtool, troubleshoot / bugfix it, then maintain thousands of packages ON TOP of designing and maintaining the distro. Let's start with the distro :)

 

See Autopackage

Link to comment
Share on other sites

hmmm...i think you are jumping ahead of yourselves a bit trying to plan out want you want in your distro, live cds and such. first you have to determine if youcan do it!

 

do the lfs thing and learn all you can first of administrating, dealing with dbs, security, etc. then when you are sure you have the time and knowledge go ahead.

 

if you don't know bash, C and C++ learn them because it will help a GREAT deal in trouble shooting when you have build errors. one thing you don't get a good idea of using distro s with packaged binaries and prebuild build scripts is just how many time you have to patch source for it to build.

 

as well packaging thing like perl packages can be a very labourous process. Or packagaing things like NVIDIA drivers. are you going to include binaries (ie flash, jre/j2sdk) or wouled you be willing to compile something likw OpenOffice. OpenOFfice is NOT easy to build at all. Lots of patches involved and a good knowledge of C++ is a huge benfit.

 

 

and so forth. .... just more devils advocacy (i do this too when friend s consider going into business for themselves)

Link to comment
Share on other sites

I figure a project plan is in order (audible groan).

Seriously.

 

I'd really start off with version <1.0 with text based install and single windows manager etc.

Keep it as slim as possible.

 

Everything else can be added later...

I agree wholeheartedly on RPM's ...

 

I also think its really important to keep standard tools that already work well. Webmin for instance ... Theres no point reinventing the wheel and personally I don't like blackbox type config aka DRAK....

 

As a minimum the tool should tell you exactly what files are being altered, preferably why!

 

 

I'm not sure about Window manager but I think something like Gnome or KDE is better than fluxbox as far as which apps get supported.

Like you use either GTK or Qt widgits and libaries.

 

On your normal PC it doesn't matter but the developers will have to handle every library etc. on the home brewed distro.

This means if you wanna use k3b then you need a whole load of KDE libaries anyway.

 

Either way, Gnome or KDE they both have lots of useful projects that can be used with less work worrying about libraries and deps.

 

What Im trying to say is:

Define a easy goal for stage 1. Make it reasonable and stick to one app per type. Preferably keep a single set of libraries and life will be much simpler.

 

Once we have a bootable and livable distro we can start incrementally adding things but if we start off all at once we'll end up drowning!

 

Break the whole project into seperate tasks and only add extras once stage 1 is working....

 

Instead of choosing favorite apps I'd define categories.

 

Windows Manager and Desktop

Packaging

Console/Xterm

Browser

Instant Messaging/Chat

CD/DVD writing

CD/MP3/Ogg/DivX Playing

Office Package

Network utilities

 

Then look how they combine using different options.

Choose the best all-round solution even if its not the favorite

Get the whole thing running/installing/tested. Get to Version 1.0

 

Once all that is done consider which individual packages should be changed out. Look at the consequences in terms of dependencies and size (if we want a live CD) I don't have a DVD-R yet!!

 

The reason I mentioned the XBox is cost. At $199 it gives a standard hardware which would be nice for a first pass at making a distro.

Like how do you test syslinux (or whatever bootable version from a SCSI CD if we don't have one.)

 

Also it would be easy to define goals for stage 1. Like multimedia, internet and office apps. It doesn't need to stop as a Xbox/Ps2 only distro but it would be good to simplify things until the team gain experience/knowledge.

 

Its more likely to succeed if each stage gives a usable product instead of try and do everything at once. The less libraries need supporting the better!

 

Ive been looking at Archlinux releases and what was in each and other than not starting with Gnome or KDE this seems to be their stategy.

Link to comment
Share on other sites

I like everything you have to say (except that xbox part still baffles me - 1) we all already have computers and 2) you're the guy with the good idea to keep it simple! And I don't have $199 for m$)

 

One thing though >

Choose the best all-round solution even if its not the favorite
Well, if it's our distro, why not pick the favorite? The "best solution" as you say is only what is best for you anyway - for example neither tyme nor I want the whole kde/gnome luggage even if they do provide all sorts of goodies.

 

So, to agree with Gowator, what if we chop down version 0.1 to

 

[*]XFree86 with Fluxbox

 

[*]Rxvt

 

[*]krusader equiv that doesn't need kde libs to be compiled

 

[*]CDBakeOven or eroaster

 

[*]xine, ogle, xmms, + a cd player

 

[*]Gaim, Mozilla firebird

 

[*]Autopackage or apt-get or rpm

 

[*]Network stuff

 

[*]Custom config GUI

 

c_m_f -> You have good ideas but KDE/Gnome distros are done to death already; I seriously believe no one can dethrone mandrake for ease of use. Newbies shouldn't use a distro created by college students ;0) We don't want to throw in everything because there's distros better than ours will be that are already like that. It's kinda like XPde.com - looks good, but then why use linux, right? Why use ours if it's exactly like the rest?

Link to comment
Share on other sites

One thing I might suggest on a web browser..Use Mozilla Firebird. Supports everything mozilla does but much lighter and faser performance..Atleast on all the machines I have installed linux on..

 

I like hte concept yall have keep it simple and light. I'd love to see a distro that doesn't focus on either Gnome or KDE.. Fluxbox I have never used but here lots of peope talk about it... Anyways I am defiantely keepin an eye on this thread as I wouldn't mind testing out a distro such as what you guys seem to want to put together..

Link to comment
Share on other sites

do the lfs thing and learn all you can first of administrating, dealing with dbs, security, etc. then when you are sure you have the time and knowledge go ahead.
I think this is great advice. I mean, if you want to do this, kiss this board, and all fun with your own pc life bye bye. You'll have no time for chit-chat and playing around if you want the first release to be out in the next 1.5 years....seriously. If you can maintain your sainity after months of using, building, and maintaining lfs, and are willing to do five times that......go for it. JMO...and best of luck to ya!
Link to comment
Share on other sites

static: i read your thread before i posted, i just thought i'd post the ideas i had at the very beginning when i decided on giving this a go.

 

ok, read all the posts, took everyone suggestions into consideration, even sarah31's warnings ;-)

 

 

plan in my mind:

  • People interested can send me an email at tyme@mandrakeusers.org and let me know your "skillz" and the like.

 

[*]Those in the core team should all do an LFS install. I plan to start mine by this weekend.

 

[*]Once the core group has completed the previous step, and is sure that they are capable of this whole thing, we'll move on to decision making.

 

now, we of course need to set down some guidelines as to the eventual plan here so that peopel can decide if they are interested or not.

 

here's the deal, in my mind:

  • slim distro, for intermediate users. requires CLI knowledge and the ability to edit configs, to an extent. not as hard as LFS, not as easy as MDK.

 

[*]try to remain gnome/kde neutral. this will take some work ;-)

 

[*]pre-made packaging system (i guess i'll give up the want to make my own *sniff*-i was basically thinking .gz's)

 

[*]hmmm...any other additions?

that list is not meant to define the exact programs we'll use...that will be decided _for sure_ at a later time.

 

stephen: GNOME is my fav, but it doesn't fit into the idea i'm looking at for this distro. fluxbox is great, I used it for quite some time.

static: i think we're pretty much on the same page. that rules.

sarah31: I know C++ pretty well, I've played with bash. I don't have any issues learning new languages, i always loved programming when I was still a comp sci major. i miss it in my new major (long story).

 

alright peeps...this is looking intriguing, but let's not get ahead of ourselves. first things first, those points i already made up above.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
 Share


×
×
  • Create New...