Re: Gnus + emacs.stackexchange

[Emanuel Berg]

Artur Malabarba <bruce.connor.am@gmail.com> writes:

>> Indeed, the participation would be needed. So start
>> with the Emacs SX! But the sx.el people already did
>> it (tho I didn't got it to work myself).
>> But because they already did it (not NNTP perhaps
>> but something to the extent of what we discuss)
>
> It's the Stackexchange API.

Well, yeah :) A solution with Gnus would be a gateway
from/to the NNTP protocol and the SX interface (or
API). Actually Gnus can do and does other things than
that (NNTP) but I don't see why anyone would choose
the small shop when there is a big. Anyway if there
already is considerable work on sx.el as it seems it
might not be a real need to have it in Gnus as well.
It would be nice tho, and it would make sense in terms
of technology.

>> I'd say the "all-in", nested thread post/comments
>> blog idea is more interesting and beneficial to the
>> Gnusers around the world...
>
> I've toyed with the idea of combining sx.el with
> some sort of mailbot, so that each question on
> Emacs.stackexchange gets sent to some list, and if
> people reply to that list the bot would post the
> reply to the site (using sx.el).
>
> Issues with this are: 1) I've no idea how (and no
> time) to write a mailbot. 2) It would have to be
> VERY smart about which replies it posts and be
> moderated. The SX API has a very incisive “Quality
> of Content” clause, which indicates the bot might
> get banned quickly if it's just posting random
> comments everywhere. 3) On SX, it's possible to
> call's someones attention by directly @naming them.
> I'm not sure how the bot would handle these
> notifications. I think the only answer would be
> someone with a lot of common sense (possibly the
> same who moderates item 2), would manually check
> all notifications.

You don't need to worry about all that. It would only
be a new *interface* to the SX sites (perhaps only the
Emacs site as a pilot case). So you don't have to
worry about doing everything SX does. You have only to
make sure hooking to SX is done in the same way it
happens when something is posted the ordinary way
(i.e. with their web GUI). If you do that then SX will
handle material the same way - how it originated (i.e.
what interface was used) will not be considered -
it'll be transparent to whatever checks and balances
that kicks in at that point.

Check out these extentions [1] to Emacs-w3m to see
what I mean. Notice they mention Google, YouTube,
different search sites like Wikipedia,
the Urban Dictionary, and so on. Notice I don't do
anything to replicate what functionality is on those
sites - I only care about hooking to them the same way
their web GUIs are. If you do that spotlessly they
have no choice but to give ou what you want. If they
don't, that means their own services are down.

[1] http://user.it.uu.se/~embe8573/conf/emacs-init/w3m/w3m-unisearch.el
    http://user.it.uu.se/~embe8573/conf/emacs-init/isbn.el

-- 
underground experts united
http://user.it.uu.se/~embe8573