From: <goran.hultgren at bluefish.se>
> "Duane Maxwell" <dmaxwell at san.rr.com> wrote:
> > From: <goran.hultgren at bluefish.se>
> > > Joern Eyrich <Joern.Eyrich at syngenio.de> wrote:
> > > > Didn't Duane Maxwell build some infrastructure for collecting class
> > comments?
> > > >
> > > > see sqdb.squeak.info
> > >
> > > Yes, I think he did and it looks nice. But it doesn't feed those
> > > comments back to the image, does it? Hmmm. I am not sure I agree with
> > > the "idea" of it. It would demand a net connection for one thing...
> >
> > Actually, I was not *requiring* a net connection any more than a Swiki
> > would.
>
> Well, what I meant was that the loop isn't "complete" yet. If I want to
> see the comments on your site I need to surf there or have some
> mechanism in the image which fetches the comment from there, which
> wouldn't work if I was unconnected.
>
> > The goal was to make it easy to collect and share class and method
comments
> > quickly - there's no restriction on feeding the entire database into an
> > image for release.  Unlike a Swiki, sqdb stores the information in a
> > structured form for easy consumption by automated tools.
> >
> > I already support multiple languages - there's an optional language tag
on
> > submissions.
> >
> > If there's anything anyone would like to see that might get something
like
> > this adopted, let me know, otherwise I'll shut it down and free the
server
> > for something else...
>
> I think that would be a waste. Hmmm. How do we close the loop then?

Well the missing piece is the image-based client to make it easy to submit
and query the database, whatever that entails - browser mods, separate tool,
harvesters, whatever.

> Could you include some form of "checkbox" somewhere on each comment so
> that harvesters could just go in, read some comments and if they/we
> think they look like a nice improvement then we check them off and your
> server bundles them all up in one single cs.gz that we could download
> and mail to Dan (or whoever is in charge of receiving the harvested
> bundles at the time). And then you would "mark" those comments as being
> "HARVESTED" or something.

Certainly, that would take but a few minutes. I just need to add the various
HARVEST fields plus a couple of authenticated moderation CGI accessors to
toggle the bits, and some minor changes to the pages.  The current pages are
meant to gently discourage web access - they work but are painful.  They
were always meant to be secondary to the XML accessors.

> If you could fix something like this - or hey, perhaps even better: make
> some easy way for harvesters to load one changeset from your site with
> all NEW comments that you have on your site so we can browse those in
> Squeak. Then we can do the gzipping/mailing from within Squeak instead.
> As a harvester it's nice to be able to work from within Squeak.

Again, a few minutes of work.  Submissions are already time-stamped (server
local time - I can switch to GMT), so *new* is easy to do.  And if I add a
separate timestamp for the setting of the HARVEST bit, then I think we don't
need the HARVESTED bit - just ask for items whose HARVEST bit got set after
a certain time, even if the submission itself is old.

> Yes, if you "build that, we will come" I think :-). Do you agree with
> me, you other harvesters?

> And btw, www.squeakdoc.org is more a readonly thing right?

Not sure - it popped up nearly the same time as mine.  It seems to have a
ghost-town quality not unlike sqdb, which got a lot of attention almost
exclusively by search engines until I blocked them.

At the time, I offered to combine sqdb into squeakdoc, but like many things
documentish, it didn't happen.  I agree it appears to be read-only, or at
least HTML-based - I know of no CGI/XMLish accessibility to its databases.
I may be wrong though - I admit I haven't paid much attention to it lately.

> > -- Duane
>
> regards, Göran
>
>