That's also what I understand, and I find this philosophy pretty appealing too, for what it's worth.
However it still worries me as this sounds a like "I'm not wearing shoes because the problem is not your feet but the pavement; What we need is a pavement that makes it possible to walk without shoes; I /could/ wear shoes until we get a better pavement, and run long distances like all those people who wear shoes, but if I start wearing shoes I will stop worrying about making a better pavement and will be stuck with a hack rather than a real solution." Furthermore, files and github are not chosen by Racket, so I don't personally mind that much using a few external tools if Racket can't do it itself (I'd still rather have Racket do it itself of course; I want a Racket machine). What's more, whether `raco` is written in Racket or in C does not make a difference, as long as it can create a package of my files. Thus, I consider `raco` an external tool. On Wed, Mar 25, 2015 at 4:13 PM, Matthias Felleisen <matth...@ccs.neu.edu> wrote: > > Thank you Greg. I couldn't have said it any better (probably worse). > This is exactly the point -- Matthias > > > > On Mar 25, 2015, at 12:07 PM, Greg Hendershott <greghendersh...@gmail.com> > wrote: > > > My personal/casual take on this: > > > > There are language systems where you to need to run some > > make-a-new-project tool -- even for a single source file. > > > > In Racket you can create multi-file collections without needing such a > > tool. Only at the point where you want to package it share with > > others, do you maybe need a tiny info.rkt. > > > > And I think the philosophy is that even this is sort of unfortunate, > > and should be as minimal as possible. For example if it were feasible > > to omit `deps` and `build-deps`, and have those automatically > > determined (not automatically genreated in info.rkt), that would be > > preferable. In many cases that would leave info.rkt containing only > > metadata like `version`. I could imagine even that instead going in > > main.rkt as a special identifier and/or submodule or whatever. > > > > So I understand the manifesto as saying, even if it turns out to be > > impractical short- or long-term, the ideal and the starting point > > should be "no external tools and config files". > > > > Also Racket has a macro system that means you probably never need to > > resort to source-file-generating/munging tools, which can be a big > > chunk of the ecosystem in some languages. > > -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to racket-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
