On Friday 17 February 2012 10:21:35 Giovanni Bajo wrote: > Il giorno 17/feb/2012, alle ore 01:02, Paul Boddie ha scritto: > > Hello, > > > > I just noticed that if I go to https://ep2011.europython.eu/, I get a > > straight redirect to https://ep2012.europython.eu/ (the front page) and > > not the archived material at https://ep2012.europython.eu/ep2011/ (its > > specific location). Would a more precise redirect not be nicer? I believe > > this was the intention when we introduced year-specific sites. > > Yes we can fix that.
Great! > We do have a problem with URLs though. The fact is that our web application > is actually a multi-year application, that is meant to host contents for > several years. Specifically, all paths under /conference are meant to be > either referring to the current year or to the previous years, within the > same website. For instance: > > https://ep2012.europython.eu/conference/speakers/paolo-sammicheli > > This is Paolo's profile page, and it makes sense "forever" the way it is > now. It will link to his talks in the different conferences, and that's it. This makes sense. If someone is to have a live profile, it's best to not let it get out of date, historical curiosity (what they were saying about themselves back in 2011) aside. > This fact (that the speaker page surives the multi-year versioning scheme) > allows you to do a multi-year navigation of EuroPython contents, like > seeing what the same speakers have talked in different years, search all > talks by tag across different years, etc. Yes, that's a good idea. > This page's URL used to be: > http://ep2011.europython.eu/conference/speakers/paolo-sammicheli > > and if you go there, you get redirected to ep2012's URL, even if Paolo has > not submitted a talk to ep2012 yet; it's just because the website > application is now under ep2012.europython.eu; thus, to preserve all > existing URLs, we had to do a redirection that preserves the path. This is > also why the root path goes to this year's root path, even though we can > special case that of course. > > There's also a few other things that you want to keep outside of the > year-versioning scheme: eg: blog / RSS feed (unless you want to lose > followers year after year). True enough. I think the aim was that you'd publish those as resources under www.europython.eu, though. I think you'd also want to publicise the profile as... http://www.europython.eu/conference/speakers/paolo-sammicheli ...although year-specific URLs would still yield the same page. > So in a way the current URL scheme is not serving us well; it's a good way > to have totally different websites each year, but it's a poor fit with our > model of keeping multiple years within the same logical website to > facilitate navigating and browsing the archives. With PyCon Italy, it used > to be: > > www.pycon.it/blog/* - multi-year blog > www.pycon.it/conference/* - archived multi-year talk content > www.pycon.it/p3/* - archived PyCon Tre (2009) website > www.pycon.it/p4/* - archived PyCon Quattro (2010) website > > but I can see that EuroPython needs to adapt to different groups managing > the website, and thus having different websites. The motivation was that you'd be able to have the current site and references available as www.europython.eu and epYYYY.europython.eu and then use the latter to archive them. The principal advantage of this is that relative links will keep working within such sites regardless of the subdomain involved, and that it gives everyone the freedom to do what they like with their sites. Unfortunately, as you point out, it raises issues about multi-year resources and it also makes linking a bit more tricky: if people link to a talk via www.europython.eu, that link will fail when the site changes for the next year. The solution to this would be mostly as you suggest: http://www.europython.eu/2011/talks/some-python-topic http://www.europython.eu/2012/events/some-python-event Then, intra-year relative links would work, inter-year relative links (still useful) would be easier, and there would be only one canonical URL for year-specific resources. > So, what we can probably do is to evolve the current ep2011 archive into > its own website (something like archive.europython.eu), that could use the > current webapp "forever". We might even decide to import previous year's > archives into it, so have a central point where all EuroPython talks are > archived (btw, were talks videorecorded during UK EPs?), and this might > make sense even if whoever succeeds us prefer a radically new website / > webapp. Also, after we import everything, we could expose the content > through a data API, so that it's easier to manage/migrate it in the future, > should we change idea again. A single site for all years, including the current one, would be best. There's practically no reason why you couldn't "mount" the different sites within a URL scheme that publishes the necessary resources as they need to be published. The only problem lies in the linking already done within those sites. If there's a URL like this... http://ep2011.europython.eu/talks/some-talk ...and some page on the same site links to it as... /talks/some-talk ...then you need to do some rewriting to get it to be... /2011/talks/some-talk As for video recordings during EuroPython in the UK, I think one of the Birmingham crowd would have to help me with my memory, but I think that at least for 2009, the recording may have been done by a non-local team. However, all recordings were uploaded to blip.tv as far as I am aware. > > P.S. Nice to see https everywhere, by the way. :-) > > Yup, I think it was due. We even got an explicit complaint about this, last > year. More https is better, especially if you can get certificates that browsers like. That was a great achievement! Paul _______________________________________________ Europython-improve mailing list Europython-improve@python.org http://mail.python.org/mailman/listinfo/europython-improve
