On 12/6/24 19:17, Lassi Kortela wrote:
>> I'm writing an article which covers (among other things) a brief
>> overview of the history of Scheme, and would like to point readers
>> towards the large body of related research.
>
> Thank you for your service to Scheme.
>
> When is your deadline?
>
> If you need a stable URL to print in a paper, we can add one under
> https://go.scheme.org/.

No precise deadline, but I intend to put it up by the end of the week.
Good to know that go.scheme.org exists, I hadn't noticed it before.

>> My current options are:
>>
>> - https://research.scheme.org/
>>
>> The nicest option, and my preferred choice, but unfortunately missing
>> a large amount of content.  (No disrespect intended, I appreciate the
>> ongoing efforts of this group and the challenges involved.)
>
> Yes, it's a matter of not enough people and not enough time.
>
> IMHO research.scheme.org is not currently in a presentable condition. If
> you'd like to improve it, that would be great.
>
> I can give you SSH access to the server. We use rsync to upload stuff.

I should have capacity to help with this over the coming weeks/months,
but need to hit a milestone or two first.

> [...]
>
>> - https://github.com/schemedoc/bibliography
>>
>> Another archive of the Bibliography and presumably the raw material
>> for the Scheme Research page.
>
> That effort started by scraping what we could salvage of ReadScheme.
>
> IIRC Amirouche (https://github.com/amirouche) copy/pasted the ReadScheme
> bibliography into Markdown format on GitHub. He also salvaged many of
> the papers (PS / PDF).
>
> That git repo is now gone, but those papers (and more) can be found at
> http://schemizdat.scheme.fi/ (user/pass "scheme"). Use `wget --mirror`
> if you like.

Very nice, thanks, I've grabbed everything.

> I started converting the Markdown into S-expressions, making quite a lot
> of progress. IIRC
> https://github.com/schemedoc/bibliography/blob/master/page9.scm is a
> complete conversion of "page9.md". Nevertheless, most of the .md files
> still have not been converted.
>
> https://scholar.google.com/ is an indispensable resource.
> https://github.com/schemedoc/bibliography/blob/master/tools/bibtex2lose.scm can parse the BibTeX citations from Google Scholar and convert to S-expressions.
>
> If you'd like to continue this conversion work, I can "match your
> contribution" as some companies say about donations. Meaning if you
> convert 10 papers, I will convert another 10.
>
> If it's too much effort for you to add the abstracts of the papers, we
> can skip that.

OK, I'll start here when I get around to this.  Perhaps some Google
Scholar scraping could be automated, eg.

https://stackoverflow.com/questions/62938110/does-google-scholar-have-an-api-available-that-we-can-use-in-our-research-applic

> [...]
>
>> Do you have any opinions or suggestions on the best way to do this?
>
> My biased (by personal interest, but also by experience) suggestion is:
>
> 1. Prefer ReadScheme on Wayback Machine now.
> 2. Start improving research.scheme.org.
> 3. Link to research.scheme.org when it's ready.
>
> If you can justify the work as part of your grant, that'd be great.

This is a good plan :)

>> Have I missed any hidden gem archive of research papers?
>
> The only hidden gems you've missed are the homepages of prominent Scheme
> scholars. E.g. Queinnec, Feeley, and Dybvig have archived a lot of their
> own stuff themselves. https://legacy.cs.indiana.edu/ is useful.

Ah, good point, thanks.  It might be worth collecting some links to
these, I have come across similar collections by Rees, Kelsey, Shivers,
Sperber, and there's surely many more.  We might as well embrace the
fact that web directories were the correct solution all along.

>> One thought I had was maybe linking to the WayBack Machine from the
>> Scheme Research page as an interim measure.
>
> That may be a good stopgap measure, but many of the links (especially to
> PDFs) are probably dead. And many of the papers only link to PostScript
> (PS) copies.
>
> research.scheme.org should provide a PDF of everything, which we can
> generate using ps2pdf if the author hasn't provided one.

Fair enough!  Once the data is all together, there will be a nice
opportunity for something like index.scheme.org to be built on top, too.