[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [hackdaworld] Read carefully this please

hi Demian

first of all, sorry for the very late answer, but i've been demotivated the last days/months.

the last time i was doing sth for hdw... the hdw structure wasn't so nice, this might be caused by base implementations... alike, "in architectures and software.. things move around the implemented base".

i agree.

i am not bothering, and i am not a native hdw developer.. and i ve been apart of this stuff during several months, but what i think is that hdw implemented base was done still with testing purposes. at least last time i left hdw we discused this. hdw needs a new implementation and may cause a different version (cause of substantial changes if these are given).

don't worry, you're never bothering. even if you do not consider yourself as a native hdw developer, you have been the only one developing for a long time and a great help.

you're right with the new version. my further plans so far:
- get installdisks working (i know you did sth like that once, i would appreciate any help!)
> no installer needed by now, you still need to be an advanced user installing hdw using hdw-get.
> fix some bugs in hdw-get.
> minimal kernel and rootfs + hdw binaries on cd.
- finish/fix the setup.d routines
> packages need to get adapted.
> hdw-get should look for setup.d entries when a package gets installed. (add hdw-get update/setup option)
- crossports (what a mess)
> booohooo, nothing works here i guess, lets fix it so that we at least get a kernel crosscompiler)
> reconsider the tc-* (toolchain) stuff hacked into hdw to ease crossbuilds (it doesn't!)
- update / rebuild all targets

having done this, we should release 0.3 (stable). after that we can continue 1.0 (devel) work, but as you said, we schould rewrite everything from scratch reconsidering and utilizing all our knowledge and experiences.

ive learned that "implemented base" in architectures and massive software is the one which defines complexity in the project. conforming project grows, added stuff is related around of the installed base. and therefore is recommendable (cause of possibly large changes) put new stuff around installed base instead of modifying this one to get rest of stuff working propertly too.


those are strongly my main purposes with the development of a new hdw version. i dont want to see hdw left behind cause a new idea wasnt incorporated cause base implementation didnt allow this. with this i do not mean the current state of hdw is bad. instead, i see beyond.

okay, stop what you learn in computer science! %) we are hackers, we run vi and start coding ... :)
all jokes aside, i think i know what you mean. that's why we should have heavy discussions !before! and during the first submitted code pieces to hdw 1.0. but basically i agree with your opinion.
imho there are some things we should consider carefully hacking on 1.0:
- flexible, extendable
- clean code, easy/simple code
(that's just what i wanted to stress out, the main things were already listed by you)
this is easy to shout out, hard to implement .. but hey, that's what we like.

other thing preocupate me is the amount of developers now working in hdw. i know that most of people who want to help and colaborate with hdw can't dedicate full time to this.. and at least, package maintenance is a hard work.. and this is my main focus when i say this. i thing that if people want to see hdw up-to-day and satisfacing needs .. many hands should be used to make this possible. if u support me... i'll try to get interested people to support hdw.

concerning package maintainance i like the way of scripts/Update-Tree. reason: people can check in changes by their own, not having access to the main scripts, which imho just one or only a few people should have. (do you agree with that?)
i could bet that if we have installdisks and released a clean, flexible and easy to improve 1.0 version more people maybe want to join hdw.

feel free to critic those aspects:
* spend some time making an analysis of current problems and proposes.
* list shared characteristics of packages hdw will work on
* describe and discuss users, different hdw targets and their promotion
* patching and developers source code management in the project
* documentation. new users and developers must see this.
* code modulation. that what we structured in scripts: "Download","Puzzle","Build".. you know.
* complete packages management discussion. this might be the most tedious thing here. how to distribute packages configuration files.
* hdw build settings. previous build procedures, build procedures, post build procedures.
* hdw install settings. previous install procedures, install procedures, post install procedures.
* build settings. how to build hdw. a cool way to make build powerful and efficient.
* build stages.
* upgrade support. most people like to see systems up-to-date.
* system configuration. support users and developers a nice interface to simplify system administration.
** there are a lot of other aspects.. i can't remember all of them now.. but i know they will surge with time.
remember i am just another voice and opinion.
i am not saying "good bye hdw"... i am saying "hello hdw.. u will stay a long time with people".

i know, and i appreciate that.

those are aspects i suggest to analize. i am not suggesting now my opinions about implementation, not until things will be discussed.
as i said before with my previous example.. the fastest and cheap solution is to ignore new object models that comes and that things continue as they are.

yes, we should keep these in mind when doing 1.0 work!

the only thing i ask for is to notify me what u think clearly. i dont ask for anything else.

i hope my plans were presented in a clear way, and even more important .. what do you think about that? did i understand you correctly?

excuse me for the talk if i am bothering. just giving my opinion as the hdw list if for i guess.

thanks for that!



If you wish to unsubscribe from this mailinglist, send mail to
minimalist@hackdaworld.dyndns.org with a subject of:
	unsubscribe hackdaworld