"On Friday 18th April the Consortium of Research Libraries (CURL) celebrated its 25th anniversary and launched it new organisational title: Research Libraries UK (RLUK). A warm welcome to RLUK and best wishes for the next 25 years!"Congratulations to them... well, maybe. I had a quick look for some key documents; here's a URL I forwarded to my colleagues a year or so ago: http://www.curl.ac.uk/about/E-ResearchNeedsAnalysisRevised.pdf. Or, more recently, try something on their important HEFCE UKRDS Shared Services Study: http://www.curl.ac.uk/Presentations/Manchester%20November%2007/SykesHEFCEStudy2.pdf. Both give me a big fat "Page not found". In the latter case, when I find their tiny search box, and search for UKRDS, I get "Your search yielded no results".
I am really, desperately sad about this. Remember all the fuss about URNS? Remember all we used to say about persistent IDs? Remember "Cool URIs don't change"? The message is, persistent URIs require commitment; they require care. They don't require a huge amount of effort (its simply a redirection table, after all). But libraries should be in the forefront of making this work. I have emailed RLUK, without response so far. Come on guys, this is IMPORTANT!
Oh and just in case you think this is isolated, try looking for that really important, seminal archiving report, referenced everywhere at http://www.rlg.org/ArchTF/. I had something to do with the RLG merger into OCLC that caused that particular snafu, and after making my feelings known have been told that "We're taking steps to address not only DigiNews links but those of other pages that are still referred to from other sites and personal bookmarks". The sad thing about that particular report, you might discover, is that it doesn't appear to be archived on the Wayback machine, either, I suspect because it had a ftp URL.
[UPDATE The report does now appear on the OCLC website at http://www.oclc.org/programs/ourwork/past/digpresstudy/default.htm. When I first searched for Waters Garrett from the OCLC home page a few weeks ago, I couldn't find it. I guess they haven't quite got round to building the redirection table yet... but that can take time.]
Grump!
I share your grump. It is somewhat depressing that this continues to be an issue. I remember feeling, 10 years or so ago, that the message about URL persistence was so self-evident that it was only a few recalcitrant, slapdash webmasters that needed to be reminded of it.
ReplyDeleteTime has proved me wrong. Even those I have some responsibility for have been difficult to maintain. We've managed it, just about, with NDAD for 10 years but it hasn't been easy. But ULCC staff pages, which I naively thought permanent enough to quote in journal articles and software documentation as a primary means of contact, disappeared pretty much without warning some years ago. I still don't understand why, nor even whether someone made an explicit decision to do it or whether it 'just happened.'
As you say, Chris, it's not fundamentally difficult to do - it just requires the will to do it. So why isn't it done, even by organisations such as libraries and archives that should be setting an example to us all ? Some time ago, I felt it was because many organisations saw the web as a mechanism for distributing ephemera - press releases, internal memos about temporary lift closures, that sort of thing - and therefore could not take permanence seriously. (That was certainly true of many government web sites in the mid-to-late 1990s, which were typically run by press officers.) But that's no longer true, and I'm not sure my original analysis was right anyway.
We do need to understand more about why this happens. I confess I don't.
I carried out a case study for ERPANET back in 2003 on the Netherlands Historical Data Archive (NHDA). At the time the future of the NHDA was uncertain as their parent institute (NIWI-KNAW) was undergoing some restructuring. Looking for their website now, it seems that the NHDA was disbanded during the course of that restructuring. Sadly, none of the links have survived - not even with a re-direct. It's quite ironic, given the nature of the institution.
ReplyDeleteI wonder if persistent identification is a common feature of web management policies? Not that policy alone would be enough to address the issue, but it would perhaps be a useful start.
It's interesting to read through the document that Chris refers to on cool URIs - in short, total control leads to stability here. The RLUK site has gone through a major overhaul that has involved moving from public sector to commercial hosting (technology lag in HE?), and its underpinning technology is more sophisticated and offers many more features (CRM, project management, and so forth - in an effort in fact to beat the very ephemeral nature of things, where sector-wide projects and initiatives often have necessarily limited funding, and short organizational life puts ad-hoc web outputs at even greater risk). The points raised are very important (and we have since worked on improving the redirect table, and indexing, for example) but amongst the “millions of reasons in practice” can I suppose be counted organisational ones – as the other examples given attest: and in terms of size, intrinsic relationship to the control question at hand, and resources, well, some organisations are more equal than others – but still find it difficult to do. In the past, publishers may have perished but their publications remained; now, with newer technology, no less, there may be a much closer relationship between persistence of organisation and ultimate availability of output (after all, it's all still there, as is RLUK).
ReplyDeleteMike, I'm pleased that progress is being made, and I do appreciate the difficulties for small organisations. However, as Kevin suggests, bodies like CURL/RLUK ought to be setting an example here. It's also going to cost you more to fix this than to do it right in the first place!
ReplyDeleteLooking at the site, I do see progress. More documents are visible, and the search screens visible on the landing page are a good idea. I guess I had expected that the URLs I particularly mentioned would have made it to the re-direct table, but an experiment suggests this is not so.
In fact, let's document this experience. I click on the CURL e-Research Needs Analysis URL (which I believe you were advertising some time ago). I get a landing page (http://www.rluk.ac.uk/curl_to_rluk), which tells me twice that CURL becomes RLUK, the second time asking me to update my links. I can report a broken link (haven't done this yet, but will do so). There is a search box, yeah! Stick in "e-Research needs analysis" (without the quotes, obviously), and hit the Search button. Uh-oh, what's this? I'm now on http://www.rluk.ac.uk/search/node, and I can see no sign of search results. Sigh, stick it in the new search box, once again. Aha, this looks like the same page but isn't, my web page is deep enough to see there are some results (telling me at the top of the page would be a good idea!). Look down the results... the document I want is not there, but there is apparently a "book page" called "Joint RLUK/SCONUL e-research Group", with some text that obviously refers to the document I want. OK I'll click on that... Aha, there in the list of documents is the one I want. Would you believe, its URL is http://www.rluk.ac.uk/files/E-ResearchNeedsAnalysisRevised.pdf. So the document does exist, and the redirection table entry in this case is simple (replace CURL.AC.UK/about/ with RLUK.AC.UK/files/).
But Mike, that's a LOT of work to make me do, and many won't do it. You're on the right track, but not quite there yet. Keep it up!