Email list hosting service & mailing list manager

remaining steps Arthur A. Gleckler (13 Dec 2020 20:16 UTC)
Re: remaining steps Lassi Kortela (17 Dec 2020 21:11 UTC)
Re: remaining steps Arthur A. Gleckler (18 Dec 2020 00:19 UTC)
Re: remaining steps Lassi Kortela (18 Dec 2020 00:50 UTC)
Re: remaining steps Alaric Snell-Pym (18 Dec 2020 07:37 UTC)
Re: remaining steps Lassi Kortela (18 Dec 2020 10:47 UTC)

Re: remaining steps Lassi Kortela 18 Dec 2020 10:47 UTC

> I, too, have been a bit useless these past few weeks (family demands,
> alas) but I should be able to dedicate a day to this project sometime
> over the Christmas break / first week of Jan, if consensus can be found
> on what I should do with that day :-)

Sounds great!

> My skills are *not* in the graphical design end, mind, but I can usually
> produce HTML/CSS that's not too horrible for human eyes, as long as it's
> plain and simple. I'm far better at writing backend Scheme, and English
> copy text :-)

Also sounds good.

>> Personally, my first priority would be the API server. If that was
>> running so that it can periodically fetch URLs and git repos, run filter
>> chains based on those, and serve the results, it would be easy to add
>> more data piecemeal, query it from JavaScript, generate static HTML
>> based on it, etc.
>
> I could give that a try. Am I right in believing we have a server to run
> it on, and it just needs a Scheme web app written to do dynamic parts
> and a cron job written to do update-cached-file parts? Can I write it in
> Chicken so I don't need to learn a new implementation so can focus on
> actually writing code - or will somebody else get started on it before
> me so I can just extend their work, which is a lot easier to do on an
> unfamiliar implementation than starting afresh as the foundation will
> already be laid? What's the list of things it should do, in priority
> order so I can do as many as I can fit into a day? Would anybody like to
> join me on the day (if schedules align) to pair on it?

I promised Peter Bex last year that we'd do it in Chicken, which has
good libraries and in particular, good web libraries written by him. By
luck, you are also most familiar with Chicken.

I wrote some draft code yesterday, here's the repo:
<https://github.com/schemedoc/api.scheme.org>. Look in the "utilities"
directory. The "src" directory has some prototype code Frank Ruben wrote
last year for Guile and Gauche, which we should integrate.

There's no fundamental block to writing the API (including the GraphQL
front-end, for which we have a working packrat parser). It's just
boring: run subprocesses, parse formats, etc. Hence I'd love to write it
together with one or more people. I can pair on it via IRC for example,
but the main thing would be just to get a steady trickle of commits
going instead of slow fits and spurts here and there :) If we get into a
gear were we constantly do something the momentum will take care of
itself. The main thing is to consciously admit if things are boring (the
opposite of what we were told to do in school for 1-2 decades) and use
collaboration to reduce that boredom.