Archive mind

[j4]

This isn't really a very Fridayish post, and as I'm going offline for the weekend in about 10 minutes (we're still internetless at home) it's going to look like a bit of a drop-it-and-run thing; but for anybody with any kind of interest in the web, a question:

You need to 'archive' some of the regularly updated bits of information which are published on various areas of your website for up to two years (no, really, you do, whether you like it or not). Your users need to be able to access the archived stuff in the same way that they can access any other bit of the website, but they probably won't actually look at it all that often, because they'll be more interested in the shiny new stuff. The info will all be backed up and archived off the web anyway, wherever (or indeed whether) you put it on the web, but you have to put it on the web and make it available to your users because that's what the nice people who give you money have asked for. 8-)

Do you

a) have a separate 'archive' section where archived material is stored, whether with this section
i) mirroring the structure of the main/current website or
ii) being organised in some way more suitable to the archive; or

b) just have a de facto 'archive' where each relevant area has the most current addition/version/whatever but also all the necessary back issues/versions/whatevers, i.e. your 'archive' material is not stored separately but integrated into the main site structure...?

I can think of various pros and cons for both (from both user and site-maintainer POV), but I want to avoid prejudicing my readers' decisions.

BTW, this is a purely hypothetical, abstract, academic question about some interesting information management issues, and not a "do my homework research before Monday's meeting" kind of question at all, oh no.

TIA as always for your input...

NB edited to add a bit about access/backups above

Comments

[nja.livejournal.com]

Depends to some extent on whether you want people to be able to see the archived material - e.g. old newletters (yes, probably, so plan b), versus old sets of safety regulations (no, probably, so plan a).

Who needs to access the archived material, and how often? Presumably the web server's backed up every now and then, could that serve as your archive or the process be adapted to serve as the archive (e.g. keeping all backups rather than just the most recent few)?

[ewx.livejournal.com]

I like stable URLs. I don't have to chase broken links that way, or find that a useful URL I quoted a week ago has now broken, and I don't end up getting Google hits that turn out not to work when I follow them. So, I don't like things being moved to an archive - I'd rather their initial location was equal to their final location. So I'd like a 'b', please, Bob.

[emperor]

If the hope is that this would be vaguely useful for your users, I'd go with "b"; that way if I remember "ooh, j4.com had a really cool article on badgers playing water polo" I can just go to j4.com and click on the "sports" link, rather than having to think "Hm, is it old enough to be archived by now, or not?".

[j4.livejournal.com]

ETA answers to your questions just before I go home! (Yes, people need to see it; people need to be able to access it all the time, even if it's rubbish [i.e. it's not a real 'archive', it's 'old stuff on the web']; everything's backed up anyway.)

HTH :) Thanks!

[geekette8.livejournal.com]

IAWTC.

[(Anonymous)]

My first reaction is to go with b but it depends on context of course.

[jvvw.livejournal.com]

Sorry that was me - didn't realise I wasn't logged on!

[caramel-betty.livejournal.com]

I'm not entirely certain what the archive would contain.

For example, say you have a list of events that you're going to be speaking at/attending/giving away freebies at. One page would want to have the current speaking events, and having the historical archive would be annoying - except, perhaps, for "We've done three out of this series of five events so far, 2 left." But then what would the archive have? It couldn't easily be left in place, since you'd want people to be able to bookmark the current events page, I imagine. So you'd need a separate archive page for that, wouldn't you?

That sort of time sensitive thing causes Issues (tm).

[nja.livejournal.com]

Oh, plan b then. I do it with the second year programming course, because it's useful for them to see the previous year's exercises and solutions (I used to allow them to see the past 5-6 years, but the lazy sod colleague who is teaching it now just recycles the exercises every year from two years previously). So to deal with Richard's objection, the current information is at the same URL all the time, and when it's updated you dump the old information into a new archive subdirectory, and replace it with the new information and a link to the new archive (and all the older archives, according to taste).

[covertmusic.livejournal.com]

b), because stuff moving around under my bookmarks/del.icio.us links/google really winds me up. :)

Shiny new stuff: RSS?

[mair_in_grenderich]

have a working search box on the site, and then I don't care how the material is organised

[imc.livejournal.com]

Other people have said what I would (leave stuff where it is as much as possible, but if it's a constantly updating page then you can save the earlier versions elsewhere but try not to move them from then on) so here's a tangentially relevant anecdote.

In 1999 Comlab radically changed the organisation of its web site, so what they did was (a) freeze "www.comlab" and start building "web.comlab" in the new structure, then at some point (b) rename "www.comlab" to oldwww.comlab and start redirecting "www.comlab" to "web.comlab". Then they changed the ErrorDocument for web.comlab to check whether the thing you were looking for was on oldwww.comlab and if so to return a message saying something of the form "sorry it's not here but you might find it on our old web site at the following URL".

This worked OK until Comlab received a letter from an administrator of the Queen's Awards saying that they'd found an expired award displayed on one of our pages (I'm not sure if they even specified which one) so please remove it immediately — at which point Comlab's administrator decreed that (rather than just remove the award images) the whole oldwww site should be removed from public view. So for a fair while, people looking for old stuff got a message which said "try this URL" and promptly got 403 when they clicked on it.

[barnacle.livejournal.com]

The items themselves could have a stable URL: the notion of an archive could merely affect how you'd navigate to them if you were coming to the site from the first time (a checkbox labelled "also search in archive" that one might uncheck to reduce the amount of long-dead noise in the results, or explicitly clicking an "Archive" link).

[barnacle.livejournal.com]

Our useability experience at work has been that long-time web-users are generally happier with either solution. The expert user will probably cope best with (b), because they'll have experience of trawling through lots of results and will know how to mentally grey out extra functionality on a page.

However, inexperienced or occasional web-users will want sections (e.g. news or press releases) to have a finite and quite small number of articles in them. They're intimidated by the idea of Google's boast of "Page 1-10 of approximately far more results than you can ever read through." Too much content, even potentially and in practice not seen, will make them uneasy with the site. Because they're working from non-computer analogies they will expect an archive to be in a place to which you navigate separately. The articles, as I mention above, need never change URL. They just need their links to potentially disappear from listings, searches etc.

(b) is what I always try to force. (a) is what our clients typically end up demanding.

[j4.livejournal.com]

you'd want people to be able to bookmark the current events page

I was sort of thinking of the events page having current events and then a "previous events" link at the end... I'm assuming some sort of date-based URL (www.thingy.ac.uk/events/2006/exciting_event.html) which could be linked to from the 'current' page or the 'archive' page. Does that make sense?

[j4.livejournal.com]

You'd start to care how it was organised when you took over managing the website!

Also, yes, some users always dive straight for the search box, but for others that's a last resort. It would be nice if the browsers/navigators could find what they wanted as well as the searchers.