Hello,
can’t we add a link checker to RAA that from time to time verifies
that a site is still alive ?
And whats about a mirror service, maybe a simple running wget so that
we still have access to the projects data even if the project owner
died.
–
Best regards,
Lothar mailto:mailinglists@scriptolutions.com
I’d be willing to volunteer server space and bandwidth for a mirror
service.
Kirk Haines
On Tue, 16 Mar 2004, Lothar Scholz wrote:
And whats about a mirror service, maybe a simple running wget so that
we still have access to the projects data even if the project owner
died.
Hi,
From: “Lothar Scholz” mailinglists@scriptolutions.com
Sent: Tuesday, March 16, 2004 6:20 PM
can’t we add a link checker to RAA that from time to time verifies
that a site is still alive ?
Interesting. Could be added the feature to RAA after replacing
current implementation (this replace is for RAA DB’s reliability
so I don’t have a plan to add new features… it’s a plan though).
When a robot detected site lost, what should RAA do?
And whats about a mirror service, maybe a simple running wget so that
we still have access to the projects data even if the project owner
http://freepan.org/ project is the one though I don’t know the
project status.
Regards,
// NaHi
Sorry about the lack of quotes. I deleted what I was going to reply to.
Ha.
On the issue of an RAA mirror, my interpretation was that Lothar was
talking about mirroring the packages themselves. i.e. following the
where-to-download links from each of the projects and pulling a mirror of
those items. That way, should they become unavailable from the original
site, they are still available someplace.
Now, on the issue of what RAA should do if it detects that a
where-to-get-it URL no longer works, assuming there is a mirror available,
what would be neat, I think, is if RAA could redirect the download link to
the mirror while keeping the original URL in the database. If, at a later
date, the original URL started working again, then RAA would simply point
back to it instead of pointing at the mirror.
That way the original hosting site, whether it be RubyForge or someone’s
personal site, is the main goto point for packages, but if anything goes
down, there is a redundant backup that can deliver the packages.
Kirk Haines
On Tue, 16 Mar 2004, Kirk Haines wrote:
Nahi, if you’re watching…
I could do http://raa.rubycentral.org and/or http://raa.rubycentral.org.
The advantage is that they would be easy for people to remember even if
ruby-lang.org is down and there are no links.
Chad
Hello Kirk,
Wednesday, March 17, 2004, 4:43:14 PM, you wrote:
Sorry about the lack of quotes. I deleted what I was going to reply to.
Ha.
On the issue of an RAA mirror, my interpretation was that Lothar was
talking about mirroring the packages themselves. i.e. following the
where-to-download links from each of the projects and pulling a mirror of
those items. That way, should they become unavailable from the original
site, they are still available someplace.
Right this was my intention. It was pure coincidence that at the day
of my posting the RAA website was also down.
And i think it is not very encouraging people who look at ruby when
the main application archive has more then 50% (that my result - looking
for email apps or libraries) garbage.
Writing a tool that periodically downloads .tgz files from the given
url and records the alive state of the link (last time failure,
download rate etc.) like every simple shareware site does shouldn't be
a problem for a software developer community.
Maybe a good way is to use the ASP (Association of shareware
professionals) XML based "PAD" format for RAA entities.
At www.asp-shareware.org there is also a toolkit for setting up
something like www.download.com in a few days.
--
Best regards,
Lothar mailto:mailinglists@scriptolutions.com
NAKAMURA, Hiroshi wrote:
When a robot detected site lost, what should RAA do?
Email the contact person(s) for the project, and after
finding the site down for 24 hours, add an “unavailable”
indication to the RAA page for the project.
On Wed, 17 Mar 2004, Chad Fowler wrote:
or how about http://raa.rubycentral.org? 
(I meant to include http://raa.rubygarden.org).
Chad
Chad Fowler wrote:
I could do http://raa.rubycentral.org and/or http://raa.rubycentral.org.
The advantage is that they would be easy for people to remember even if
ruby-lang.org is down and there are no links.
or how about http://raa.rubycentral.org?
(I meant to include http://raa.rubygarden.org).
I vote for all three. 
Actually, might it not be a good idea for things to migrate
gradually to rubyforge?
AFAIK you don’t have to use “all” its features just to use it
for storage.
Hal
Hi,
From: “Chad Fowler” chad@chadfowler.com
Sent: Wednesday, March 17, 2004 4:52 AM
Nahi, if you’re watching…
I could do http://raa.rubycentral.org and/or http://raa.rubycentral.org.
The advantage is that they would be easy for people to remember even if
ruby-lang.org is down and there are no links.
or how about http://raa.rubycentral.org?
(I meant to include http://raa.rubygarden.org).
http://ruby.yi.org/raa/ is the only mirror of RAA for now.
The site retrieves RAA information via SOAP.
This kind of read-only mirror is not hard to setup I think.
Regards,
// NaHi
On Wed, 17 Mar 2004, Hal Fulton wrote:
That’s a mighty progressive though. If we’re (the community) ready for
it, I think it would be fantastic. No need for a separate RAA and
RubyForge when RubyForge could be RAA.succ.succ.
What do the RAA and RubyForge guys think?
Chad
Chad Fowler wrote:
Chad Fowler wrote:
> # I could do http://raa.rubycentral.org and/or http://raa.rubycentral.org.
> # The advantage is that they would be easy for people to remember even if
> # ruby-lang.org is down and there are no links.
>
>
> or how about http://raa.rubycentral.org?
>
> (I meant to include http://raa.rubygarden.org).
I vote for all three.
Actually, might it not be a good idea for things to migrate
gradually to rubyforge?
AFAIK you don’t have to use “all” its features just to use it
for storage.
That’s a mighty progressive though. If we’re (the community) ready for
it, I think it would be fantastic. No need for a separate RAA and
RubyForge when RubyForge could be RAA.succ.succ.What do the RAA and RubyForge guys think?
Would it be helpful to have a RAA config (which I knov nothing about, I
hasten to add) that could be committed to?
I wonder if that would make keeping links updated simpler?
Now rurby stdlic doc has matured, and RubyGems seems off to a flying
start, it would be great to see a fledging CPAN-a-like start off too.
All this and the Poignant Guide too…
it’s been an interesing year so far 
On Wed, 17 Mar 2004, Hal Fulton wrote:
Chad Fowler wrote:
Actually, might it not be a good idea for things to migrate
gradually to rubyforge?
AFAIK you don’t have to use “all” its features just to use it
for storage.
That’s a mighty progressive though. If we’re (the community) ready for
it, I think it would be fantastic. No need for a separate RAA and
RubyForge when RubyForge could be RAA.succ.succ.What do the RAA and RubyForge guys think?
Actually, that’s not quite what I meant.
RAA has always been about metadata, and Rubyforge is about hosting.
I’m just talking about increasing the tie between them – encourage
every RAA project owner to host on RF.
They’d still be separate entities that way, but the centrality of
Rubyforge would mean no broken links unless that server went down.
Hal
On Wed, 17 Mar 2004, Hal Fulton wrote:
Hal Fulton wrote:
RAA has always been about metadata, and Rubyforge is about hosting.
I’m just talking about increasing the tie between them – encourage
every RAA project owner to host on RF.They’d still be separate entities that way, but the centrality of
Rubyforge would mean no broken links unless that server went down.
Perhaps we can get RubyForge to export a metadata file (in, say XFML, or
XTM , or maybe some RDF thing, something that can transport the basic
info, is easily aggregated, plays well with XPath/XQuery, etc.) that
RAA understands.
If you host an app on RubyForge, then you need never maintain anything
on RAA; the metadata feed has your project info.
If you prefer to host someplace else, then give RAA a link to your own
metadata feed.
If that doesn’t work for you, then maintain the metadata directly in the
RAA.
James
In article 40577CFD.9060702@hellooperator.net,
Chad Fowler wrote:
Chad Fowler wrote:
> # I could do http://raa.rubycentral.org and/or
> # The advantage is that they would be easy for people to remember even if
> # ruby-lang.org is down and there are no links.
>
>
> or how about http://raa.rubycentral.org?
>
> (I meant to include http://raa.rubygarden.org).
I vote for all three.
Actually, might it not be a good idea for things to migrate
gradually to rubyforge?
AFAIK you don’t have to use “all” its features just to use it
for storage.
That’s a mighty progressive though. If we’re (the community) ready for
it, I think it would be fantastic. No need for a separate RAA and
RubyForge when RubyForge could be RAA.succ.succ.What do the RAA and RubyForge guys think?
(for some reason the parent of this post didn’t show up in my newsfeed, so
I’m replying to this one)
(Ok, you didn’t ask me, I’m in neither group mentioned, but I’ll put in
my 2cents anyway)
Why not? This is a golden opportunity to see some real change since the
RAA is in maintenance mode till the end of the month. It’s worth a try.
Lots of folks in the community have been waiting for RAA.succ for 2-3
years now. Now is not the time to be timid. 
If it works out, it’ll be great for the community all around. If not,
then maybe it’s still a bit too early and the two (RAA and
RubyForge_with_Gems) will coexist for some amount of time till one wins
out over the other based on user support (I would think that
RubyForge_with_Gems would eventually win because of the convenience, but
who knows?)
“To ask permission is to seek denial” – Usually attributed to Scott
McNeally, but I’m sure he was quoting someone.
Phil