What would it take to get trn on the web?

[nancylebov]

It seems to be technically possible now that Javascript is in common use. It would make long discussions easier to read and navigate. I believe that making long discussions more feasible would make the web a better place, and killfiles might be a noticeable win too.

I've mentioned this, and been told that it wouldn't be that hard to write, but it's too boring.

The question has been raised of whether money would make it interesting, and if so, how much. Any opinions?

(I put this in terms of trn because that's what I was happy with. I don't know whether slrn has a huge advantage, or if it would be much more trouble to write.)

Comments

[autopope.livejournal.com]

slrn has a huge advantage over trn.

First there was rn; with no threading, but killfiles.

Then there was trn, with threading -- and it was basically fit for purpose.

Then there was strn, with threading and scoring, so that rather than using an all-or-nothing killfile to block unwanted content you could use a nuanced score file to raise or lower the interestingness of a given post.

slrn is a from-scratch rewrite with the functionality of strn but a CURSES interface reminiscent of tin (but vastly more powerful). Interestingly, it uses a macro language and graphics library, S-LANG, mostly as an abstraction layer between the newsreader and the terminal.

[nancylebov.livejournal.com]

Supposing that someone wanted to write a web version, how much harder or more time-consuming would it be than trn?

[autopope.livejournal.com]

Either way, you're proposing rewriting a large chunk of software -- several thousand or tens of thousands of lines of C -- in JavaScript, a radically different language. There's also the question of JavaScript's suitability for low-level TCP/IP communication as opposed to high-level HTTP-mediated message passing, and the question of what kind of server to run at the other end.

Certainly web forum designers ought to be beaten with a dead haddock until they study NNTP and NNTP newsreaders exhaustively before they reinvent the wheel. But the real solution ought to be something NNTP-sever-like (but more spam-resistant) running on the server, and talking to a browser-hosted JavaScript app via AJAX; an app using HTML 5's local file storage/database extensions to store killfile/scorefile information locally, and so on.

But that's an enormous job.

In fact, if I was looking for a business start-up, I'd be drawing schemas on the back of a napkin and looking for angel investors. Because you're looking at more than one programmer-year of work to do it right.

[nancylebov.livejournal.com]

Thank you. That was an important piece that I just didn't have a feeling for. I could tell there were a lot of details which would need to be right, but not how hard the underlying task was.

I was imagining it as shareware, but I suppose companies which have commercial forum software would pay for it.

I bet that if there was a commercial version, a shareware version would materialize soon enough.