Sustainability of purl.org

[Norman Gray]

Greetings, all.

Does any know about, or can anyone point to a discussion of, the longer term sustainability of the OCLC PURL service at purl.org?

(I'm aware that this list is really about the purlz software rather than OCLC -- I'm asking for community gossip rather than authoritative statements, and this seems a reasonable place to ask.)

A couple of times recently, the purl.org has been at least briefly down or not responding, I have the impression that the purlz version onpurl.org is a little behind the released version, and have the impression that admin requests requiring human intervention can occasionally take a little while for a response. That is, purl.org isn't not-working, but it has a rather ... unloved air.

To be clear: I am NOT complaining here -- I know that OCLC contribute purl.org as a service to the community (plaudits for OCLC!), that it's supported on a best-efforts basis, and that I'm not paying anything for the service. That said, purl.org is quite important to some bits of the web, and one naturally wonders whether one should continue to advocate it as a good practical solution to long-term naming of URLs.

So: does anyone know if there are long-term plans here, or if there's likely to be, for example, a move towards sustainable funding?

To be concrete: I can imagine a funding model based on requiring or requesting users to pay a small 'endowment' to register a PURL. A yearly subscription (on the model of DNS registrations) wouldn't be right, because the point is (to be able) to forget about a purl once it's registered. Calling it an 'endowment' rather than a 'fee' has the right associations.

Of course, the next question is the size of that fee/endowment. I wouldn't turn a hair at USD 2 per PURL, and would probably go up to about USD 10 without complaining too much. Or I might hand over, say, USD 20 up-front as credit in advance to be able to later register some number (3--10?) of PURLs; perhaps I would go higher. I've no idea whether those represent useful numbers, since I've no idea how many PURLs are registered per year, or how many could be registered if the service were promoted more.

Myself, I have one or two dozen PURLs registered (under http://purl.org/nxg/*). I use them as the canonical URLs for software I release. I'm not sure anyone pays attention to the PURLs, rather than perhaps bookmarking the URL they resolve to, but I at least get the warm glow.

Best wishes,

Norman


--
Norman Gray : http://nxg.me.uk

[David Wood]

Hi Norman,

On May 6, 2014, at 16:55, Norman Gray <norman...@gmail.com> wrote:

> Greetings, all.
>
> Does any know about, or can anyone point to a discussion of, the longer term sustainability of the OCLC PURL service at purl.org?
>
> (I'm aware that this list is really about the purlz software rather than OCLC -- I'm asking for community gossip rather than authoritative statements, and this seems a reasonable place to ask.)

It has been a while since I dealt directly with OCLC, but here are my impressions:

Cons:

- The purl.org operation remains in OCLC’s R&D division and has never moved to their production team AFAIK.
- You are correct that there have been some stability issues. This would seem to be because it is still deployed on a single dedicated hardware.
- OCLC appoints one part-time person to manage the site at a time.

Pros:
- The purl.org site has been running since 1995 or so.

As much as I can and do criticize OCLC for the cons, that single pro is really quite impressive. Next year will be 20 years!

The Linked Data community is one of the external stakeholders that has heavily relied upon OCLC’s largesse with purl.org. However, some people (including myself) started https://w3id.org/ last year due to the stability issues. Unfortunately, that site is implemented using Apache and GitHub instead of a PURLs implementation, but the Web doesn’t care.

Regards,
Dave
--
http://about.me/david_wood

>
> A couple of times recently, the purl.org has been at least briefly down or not responding, I have the impression that the purlz version onpurl.org is a little behind the released version, and have the impression that admin requests requiring human intervention can occasionally take a little while for a response. That is, purl.org isn't not-working, but it has a rather ... unloved air.
>
> To be clear: I am NOT complaining here -- I know that OCLC contribute purl.org as a service to the community (plaudits for OCLC!), that it's supported on a best-efforts basis, and that I'm not paying anything for the service. That said, purl.org is quite important to some bits of the web, and one naturally wonders whether one should continue to advocate it as a good practical solution to long-term naming of URLs.
>
> So: does anyone know if there are long-term plans here, or if there's likely to be, for example, a move towards sustainable funding?
>
> To be concrete: I can imagine a funding model based on requiring or requesting users to pay a small 'endowment' to register a PURL. A yearly subscription (on the model of DNS registrations) wouldn't be right, because the point is (to be able) to forget about a purl once it's registered. Calling it an 'endowment' rather than a 'fee' has the right associations.
>
> Of course, the next question is the size of that fee/endowment. I wouldn't turn a hair at USD 2 per PURL, and would probably go up to about USD 10 without complaining too much. Or I might hand over, say, USD 20 up-front as credit in advance to be able to later register some number (3--10?) of PURLs; perhaps I would go higher. I've no idea whether those represent useful numbers, since I've no idea how many PURLs are registered per year, or how many could be registered if the service were promoted more.
>
> Myself, I have one or two dozen PURLs registered (under http://purl.org/nxg/*). I use them as the canonical URLs for software I release. I'm not sure anyone pays attention to the PURLs, rather than perhaps bookmarking the URL they resolve to, but I at least get the warm glow.
>
> Best wishes,
>
> Norman
>
>
> --
> Norman Gray : http://nxg.me.uk
>

> --
> You received this message because you are subscribed to the Google Groups "persistenturls" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to persistenturl...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[Norman Gray]

David, hello.

On 2014 May 7, at 15:17, David Wood <da...@3roundstones.com> wrote:

> It has been a while since I dealt directly with OCLC, but here are my impressions:
>
> Cons:
>
> - The purl.org operation remains in OCLC's R&D division and has never moved to their production team AFAIK.
> - You are correct that there have been some stability issues. This would seem to be because it is still deployed on a single dedicated hardware.
> - OCLC appoints one part-time person to manage the site at a time.
>
> Pros:
> - The purl.org site has been running since 1995 or so.

Very much so, and yes, ~20 years is impressive. I think I registered the subdomain purl.org/nxg more than half of that time ago.

purl.org's Big Thing is the promise that it's not going away, and the technological simplicity that makes it manifestly feasible to move the identifiers to a different host if OCLC finds maintenance challenging.

I remember a workshop on persistent identifiers of a few years ago, where Stuart Weibel (I think; or it may have been John Kunze) made this point very convincingly. Something under purl.org or under id.loc.gov has an "institutional commitment to persistence" which is worth an awful lot more than any amount of indirection that you get through a fancy URI scheme. As Stuart (or whoever) said , "loc.gov isn't going away any time soon".

Thus purl.org isn't a technological solution to a problem, but a (great) social one. But the problem with that is that it works only as long as everyone believes in it, and the vague air of neglect hanging around purl.org makes that belief a little bit more difficult than it used to be. Even a thread like this, openly speculating about purl.org's sustainability, might be seen as contributing to the confidence-rot (and I delayed asking the question for a long time, for precisely that reason).

> The Linked Data community is one of the external stakeholders that has heavily relied upon OCLC's largesse with purl.org. However, some people (including myself) started https://w3id.org/ last year due to the stability issues.

That's a nice solution.

So should 'everyone' move to w3id.org? The readme there suggests that all it takes to join in is a github pull request for with a redirect of one's choice -- is that correct? Hmm: w3id.org/nxg/... has a ring to it.

I'm rather reluctant to do that, since jumping onto bandwagons is precisely the opposite of what one should do with persistent identifiers.

But I'm now becoming equally reluctant to advocate purl.org as the best practical solution to the persistent identifier problem.

Is this the best place to have this conversation? What to do, what to do, ...? (<-- that's not a rhetorical question)

All the best,

[Norman Gray]

Greetings.

On 2014 May 10, at 23:09, Norman Gray <norman...@gmail.com> wrote:

> So should 'everyone' move to w3id.org? The readme there suggests that all it takes to join in is a github pull request for with a redirect of one's choice -- is that correct? Hmm: w3id.org/nxg/... has a ring to it.

and

> Is this the best place to have this conversation? What to do, what to do, ...? (<-- that's not a rhetorical question)

The message <http://lists.w3.org/Archives/Public/public-perma-id/2013May/0000.html> answers these questions to some extent, and I've just joined that (very quiet...) list.

I still don't want to just walk away from purl.org, though.

[David Wood]

Hi Norman,

You are correct to my mind - the social commitment is the key and that is hard to come by. However, I would personally rather trust the LoC than OCLC for the long term due to the upheavals discussed in the library community for more than the last century. Rapid technological advancement has strained libraries even harder as of late.

My preferred approach is to widen the social commitment and make it inherently multi-organizational in nature. A few of us started thinking about that a few years ago and ended up with this (unimplemented) idea:
https://code.google.com/p/persistenturls/wiki/PURLFederationArchitecture
It could be improved, but was always intended as a starting point to bring other interested parties into the mix. The goal in brief was to have multiple organizations jointly oversee a DNS domain that would at runtime resolve to any number of peers. Crucially, any single organization could come and go as they wished or needed to without jeopardizing the whole enterprise.

I had hoped to convince OCLC that they could become a “super-peer” for purl.org, but that didn’t happen. There was no appetite on their side to address the concerns, much less take any action.

I’m copying the W3C Pemanent Identifier Community Group to bring them into this discussion.

Any interested parties should to subscribe to one or both lists to ease the administrative burden of manually approving each message.

W3ID folks should review this thread:
https://groups.google.com/forum/#!topic/persistenturls/xEeXU76oewk

Regards,
Dave
--
http://about.me/david_wood