Posted by Chris Uppal-3 on Apr 27, 2004; 10:47am
URL: https://forum.world.st/What-has-happened-to-Dolphin-Harbour-tp3370442p3370454.html

Sean Malloy wrote:

You seem to be raising three separate specific issues.

1) The state of the Dolphin help/website.
2) The absence of a browseable collection of "how-to"s, or "snippets".
3) The absence of a clear way to find existing resources.

Regarding (1), I have no comment.  That's up to OA.

Regarding (2):

> You ask a question, you get an answer
> (most of the time more than one). I'm saying perhaps some Dolphin
> developers should be more pro-active, and publish information, whether
> someone has asked or not. Know what I mean?

As a regular poster, and a poster of snippets at that, I'll explain why I,
personally, would have no interest in contributing to such a collection.  Not
that I'm saying the idea is bad, only that I wouldn't bother adding to it.

For me, this newsgroup is primarily about discussion.  I read it because it is
/interesting/, not because I have some selfless drive to "help newbies and
promote the Dolphin cause". (I'm quite happy to do either of those things, but
they're not the reason I'm here).

Most often, if I post an answer to a question, or a snippet of code, it because
I found the question interesting.  The typical case is that I /didn't/ know how
to do <whatever>.  So I read the question, and become interested in finding out
the answer.  Often I have an idea of how to do it, but want to check see if my
idea is right, so I test it.  Sometimes I have a less precise idea of where to
start, so I have to do some research.  Sometimes I have /no/ idea where to
start and have to do a lot of research.  (Or, of course, I may have no idea how
to do it, and no interest in finding out either -- as if often the case for
COM/ActiveX-type questions).

If I've "solved" the question (or if I have a partial solution, and no one has
posted a fuller solution), then I may as well post it.  Why not ?  It's not as
if I have any fear of pontificating ;-)

The other part of this, is that I get to compare my "solution" with other
people's.  They can (and do) correct, expand, or question what I've posted.
Even if they don't do so directly, then it's always interesting to compare, and
learn from, other people's approaches (which are often very different from
mine).

There are other occasions when I'll try to post answers to questions (that I
already do know an answer to), but that tends to be when I think the answer
itself is interesting and worth discussing.

So, for me, the idea of adding to a published collection of snippets and
how-tos holds little appeal.  Other people, of course, will have different
opinions on this.

Incidentally, and independently, another reason I'd be disinclined to
contribute to the collection is that there is already a very large collection
of examples in the form of the source code to the system (and in any third
party stuff you add).  That is /intended/ to be browsed and learned from.  I
don't know how much you are in the habit of looking for examples and insight in
the system code, but if you are not, then you are missing probably the single
most valuable resource there is.  I've suggested to Andy before that this point
should be made clearly in very first introductory material that a new user ever
sees.

And don't forget that Ian maintains an archive of this NG.  And -- for more
recent posts -- there's also Google.

Regarding point (3) ("the absence of a clear way to find existing resources" in
case you've forgotten by now ;-)

I think this is something that is well worth thinking more about.

One way would be to have a communal "links" page somewhere central.  But that
means "somewhere on the OA site", by definition.  There used to be a "links"
page as part of the wiki, but I gather that that is now frozen (I haven't
checked).  Unfortunately, even when we could edit that page of links, very few
people did.  I suppose website creators were bashful about adding links to
their own sites to it (I know I had to *force* myself to add mine -- it would
have been much, much, easier just to let it lie).  Also there's the usual
problem of people being unwilling to change (as opposed to add to) the stuff on
a wiki page.

Another alternative is a links page that /isn't/ part of the OA site.  But we
already have such pages, and they (as you have obviously found) aren't enough.
One problem is that maintaining them is chore.  Another is that the newbie has
to /find/ them, and that's no easier than finding any other resource.

So, in the absence of further data, it appears that a centralised links list
won't work.  So how about a de-centalised one ?

The web-ring is a decentralised approach, but it doesn't seem to work too well.
I know I haven't bother to get my website on the rign, so maybe I'm part of the
problem myself.  Anyway, I didn't find the ring too appealing for some reason.
Possibly because they tend to collect dead links.  Possibly because I didn't
fancy adding javascript to my site.  Possibly because I didn't like the central
web-ring site.  Possibly because I'd prefer (as a user of a web-ring) to see
some indication of what a site has /before/ I follow the link.  I dunno.  Does
anyone think it's worth having a drive to get the web-ring healthy again ?

Or, another decentralised approach.  Someone (I'd be willing to do it myself,
if there's enough call for it, and buy-in -- in advance) would maintain "links"
list.  That list would be posted to this NG about once a month.  We, as a
community, would be responsible for helping keep the list up to date.  Several
(ideally most) of the third party web-site owners would undertake to mirror the
*current* list somewhere on their website (that's what I mean by buy-in), so
the list should be in the form of /simple/ HTML so as to make it easy for
website owners to paste into their pages.  The list should also have a "last
checked and updated at" date featured prominently somewhere -- a stale list is
of little use.

I don't know, but maybe it would be possible to create a rough index of the
packages that are available and distribute that in the same way.  I suspect
that would be going too far, and betoo difficult to maintain.  Thoughts anyone
?

Lastly, you mentioned the squeakmap in another post.  I'm not sure how that
works, but it /looks/ as if it is a centralised list of packages, plus a
feature to download them and install them directly into your image.   There are
security issues with that (I /don't/ like the idea of executing random
third-party code before I've seen it -- and loading a package into an image
does just that).  Those issues could probably be avoided by making it
browseable, perhaps using Rosetta instead of .pac files.  Again, it seems to me
that the centralised nature of the map means that it would /have/ to be hosted
(and probably lead) by OA.  Also it'd be a big project[*], and would have
considerable impact on people who contributed content.

    -- chris

[*] and it would probably use so much third-party code (e.g. Steve Waring's
HTTP stuff) that just installing it would give a pretty complete index of
what's available ;-)