wanderer wrote:
its awesome
thanks for all your work
wanderer
i really appreciate it.
i want you to know, that although im happy with the 2 years of work ive done on this sort of thing (which started as a way to remix puppy) that its not so much about the script as the philosophy. the script is just an implementation of that philosophy.
i think about a future where distros themselves are less important, and tools to create distros are aimed more and more at users rather than developers.
traditionalists will laugh at this. users cant create distros. theres a lot more to a distro than remixing an existing iso.
ive spent a decade working this philosophy out-- its largely based on my experience with puppy at least 12 years ago, and woof-ce proves theres a lot more to a distro than remixing an existing iso.
that said-- not everyone wants to build up from the core, some want to mess with coreplus instead.
we have a framework that lets us offer both.
its simple enough for people to understand. its a little unfamiliar, superficially, but a lot of the "sophistication" it lacks is stuff you can clearly tell i know how to do.
there are features where lists of files are used instead of arrays. there are features where arrays are used. there are features where strings are passed directly to the shell, too.
each time ive added something, ive tried to do it the easiest way-- not the "best" way.
both are subjective-- different people think different things are easy. different people think different things are simple-- or best.
and theres a good reason for it. even though i could refine and refine and refine-- what i really want is to make it so that people can take it and do wtf they want. i dont like the "wtfpl" license, licenses shouldnt be sloppy, but this is a "wtfyw" design.
in "better" languages like bash (bash is a better language, of course. powerful too.) a "wtfyw" design will be very difficult to manage down the road.
this is done with language thats designed more closely with a "wtfyw" (do whatever the f you want) philosophy. take all the script and put it together. take all the script and separate it into other files (if you want to.) make your own version-- use an existing version.
the only reason weve made it this far this quickly is because of your attitude and your philosophy.
it just happens to be one that i can assist. and im very grateful to you for that, because i did design this stuff in hopes of helping somebody.
you just happen to be really fun and really easy to help. if you become more familiar with what ive done with your work, youll find:
1. its surprisingly easy
2. youll find ways to make it easier
im proud of my language, as well as its simplicity, but i still think a language designed specifically for remastering gnu/linux would be really neat.
mcorepup actually began as a program to compare different puppy isos. there are hundreds (for starters) and i wanted to make a table of everything important in them.
at one point it gained the ability to produce bootable isos.
the feature that lets you say whose changes to add is new to corepup, and i think very important. it lets the developers choose a default config, and the user change that, and other people add to it without everyone saying "oh no, i dont want this change."
instead of bothering people about adding stuff, trying to "protect" it it from changes, it just makes it more trivial to prevent their changes from making it into the iso. just remove the name, and you control the contributors yourself (as the lead dev or as the user, both get a real say.)
but the philosophy is simpler than that-- put every user in far greater control. use automation so they dont have to do it all themselves.
instead of making distros-- make tools to make distros.
heavily inspired by woof itself, though olpc has a tool to turn ubuntu into their own olpc platform. we didnt invent automated remastering. but isnt it cool?
as for the future of the distro-- a generic bootable iso that isnt fun or interesting (but can be remixed the way we are doing) isnt rocket science.
what do you get if you have lots of tools for remixing every aspect of a distro?
you have a distro factory, which is aimed at users running the factory instead of these so-called "developers."
im not anti-developer. its really the user/dev dichotomy that im trying to reduce. every user should have more of a taste of developing, even if theyre lazy (most users are.)
and this philosophy would make it easier to become a developer-- it would reduce the overhead of modern large distros-- it would make organisations like spi and red hat less powerful, and thats why eventually youre going to hear people speaking very loudly against this philosophy.
but its too cool not to explore, and no matter how simple your approach (even because your approach is simple) you are a pioneer. and rockedge is a bigger deal than he probably understands.
you cant change the world with 2 people as easily as you can with 3 or more. you guys are my heroes. we are not elite, we are a few guys messing around with software in a garage, doing stuff our way.
not saying we will be famous, not saying that at all-- but apple and hp (and puppy and the linux kernel) started that way too.