1. Hi there,

   I can help you out with this, since I deal with it almost every day – and every month at our Operations meeting.

   There’s a good reason why the ‘www’ part has to be given out as the website address.

   Firstly – it only has to be read out on radio – because the website cannot actually be seen by listeners (although, it can on DAB Radio).

   However, when giving out a URL on television, presenters can drop the ‘www’ part of the URL as long as the site address appears in full on-screen.

   The reason the ‘www’ part has to be given out is exactly as you describe – if you drop the ‘www’ part from the following address:

   http://www.bbc.co.uk/scotland

   You would be redirected to the main BBC homepage at http://www.bbc.co.uk

   This doesn’t happen with addresses on Channel Four’s website because they are http://www.channel4.com – and that particular domain is not used for anything other than the Channel Four website.

   The BBC use bbc.co.uk for all manner of things – not just the public-facing website.

   The BBC internet infrastructure is also absolutely massive – it’s actually a network unto itself. There are currently around 24 BBC content servers, roughly split between Docklands in London, and Telehouse in New York.

   The reason we have servers in New York, is to cope with the demand from trans-Atlantic users, and save on expensive data transfer from London to the USA and beyond.

   With this set up, aliases are controled by the bbc.co.uk domain name server (DNS) in London.

   Remember that the BBC has other bbc.co.uk aliases – not just external ones like news.bbc.co.uk – but also internal servers for staff use, like gateway.bbc.co.uk and other development and testing servers. All of these aliases can only be controlled from one single machine.

   The domain name server also handles redirects (for example sending users from http://www.bbc.co.uk/sportsound to http://www.bbc.co.uk/scotland/sportscotland/sportsound).

   In order to drop the www part from URLs, we would have to put redirects on every single one of the servers – which can’t be done, because there can only ever be one domain name server for a domain name.

   If you had more than one – people wouldn’t be able to access your website – since the DNS is attached to the domain, and every computer in the world connected to the internet looks for the DNS of the website first, then converts it to the domain name and displays the webpage – which is why you get http://www.bbc.co.uk when you try to look for bbc.co.uk/scotland

   The alternative is splitting off the public-facing website to a different domain entirely from bbc.co.uk – and it’s probably too late in the day to do that now.

2. OK, the third last paragraph made my stomach turn for a second there …

   There’s a difference between domain name servers – servers which dish out hostname to ip-address mappings to clients, i.e. web browser, ftp client, ssh client, mail server, etc. – and Apache sitting on a server saying “hello, you asked me for /scotland, i’m passing you to http://www.bbc.co.uk/scotland“.

   If by domain name server you mean a server sitting attached to the root TLD bbc.co.uk, then saying there can be only one is untrue. You can assign many A records to a root domain, and clients will cycle through them in round-robin fashion. If you want to be uberl33t, stick a load-balancer on that IP and assign many machines to it.

   I note that when connecting to bbc.co.uk on port 80, it tells me that it’s actually www-bogus.kw.bbc.co.uk, presumably down in Kingswood Warren from the hostname. I note that it’s running Apache. It’s not running an SMTP server/daemon, so it’s not required for mail.

   I also note it gives out the old “this is not currently a public BBC server” when you connect to the IP address, and not the hostname “bbc.co.uk”. This means that there’s a VirtualHost configured on that box for bbc.co.uk, and it’s redirecting to http://www.bbc.co.uk, no matter what REQUEST_URI is being asked for from the client. This also means that it’s then trivial to change whatever code is there to add Redirect / http://www.bbc.co.uk/ to the httpd.conf.

   If it’s non-trivial, then I’m unimpressed with the otherwise superb BBC network! 🙂

   In the meantime, allowing presenters to drop “www” from web addresses, despite the on-screen presence, is just laziness both on the part of the presenters and the part of the policymakers – presumably the presenters couldn’t be arsed saying “double-yoo double-yoo double-yoo” all the time 🙂

3. OK, checked into this a bit more …

   intrepid:~> telnet bbc.co.uk 80
   Trying 132.185.132.204…
   Connected to www-bogus.kw.bbc.co.uk.
   Escape character is ‘^]’.
   GET /scotland/ HTTP/1.1
   Host: bbc.co.uk

   HTTP/1.1 302 Moved Temporarily
   Date: Thu, 21 Aug 2003 18:42:17 GMT
   Server: Apache/1.2.1
   Location: http://www.bbc.co.uk/
   Transfer-Encoding: chunked
   Content-Type: text/html

   ae
   <HTML><HEAD>
   <TITLE>302 Moved Temporarily</TITLE>
   </HEAD><BODY>
   <H1>Moved Temporarily</H1>
   The document has moved <A HREF=”http://www.bbc.co.uk/“>here</A>.<P>
   </BODY></HTML>

   That’s not good! 🙂 It’s almost doing what I’ve suggested, and the worse part of it is that it’s utterly, utterly trivial to fix. 🙁

4. Interesting that you think it’s trivial to fix.

   Another point you haven’t considered:

   when people access a BBC website, depending on where they are located – they are given any one of the main servers on the network.

   Putting in http redirect configs onto all of the possible servers doesn’t solve the problem – because we have minimum page request times and download times to consider. Adding more handlers to each individual request would increase download times.

   I don’t mean to be rude – but if your own webpage is anything to go by, the BBC would lose a lot of customers if our webpages took as long to download as yours. Opera returned a total download time of 29 seconds. If I’d held my breath, I would have probably suffocated.

   Tell you what: you provide a complete technical suggestion on how we can do this, I’ll put it to our guys and come back to you with an answer.

5. > Putting in http redirect configs onto all of the possible servers …

   I didn’t suggest this, I suggested putting http redirect configs on *one* server – bbc.co.uk – which, according to DNS, is one IP address.

   > I don’t mean to be rude …

   Well, a) my site is hosted over my DSL, so compared to the mighty BBC content network, I only have 256kbit/sec to let people load my site, b) when you load my home page, you’re loading a lot more content than the BBC page, c) I haven’t bothered optimising the code because it’s just my own personal site, and not one for a client, and d) I didn’t realise the BBC had “customers” – I’m a licence fee payer, don’t I own all your asses? 🙂

   > Tell you what: you provide a complete technical suggestion …

   Without knowing the complete layout of the content network, I wouldn’t be able to provide a complete technical suggestion. I can only operate on empirical evidence, which says that: bbc.co.uk is one box; redirecting to http://www.bbc.co.uk need not break any location dependence, since that’s what’s being done already; and that half the job is done already. If I saw the Apache config files and scripts that are doing the redirect at present, I’d be in more of a position to “suggest” more completely 🙂

6. Oh, and I’ve never seen an in-header 30x Redirect take more than a fraction of a second before – and since bbc.co.uk does this already, sans a few extra bytes for the REQUEST_URI info, where’s the issue? 🙂

7. I have a partial answer on this for you.

   bbc.co.uk isn’t an official URL. The proper URL is http://www.bbc.co.uk – which is where the DNS is attached to, along with numerous other resources.

   Therefore, bbc.co.uk can only ever be a static IP address – it can never have the same characteristics or resources as a full URL, like http://www.bbc.co.uk

   If we started allowing that IP address to handle bbc.co.uk/anything to redirect to http://www.bbc.co.uk/anything, then it would have to be able to cope with immense demand (you’ve probably got no idea just how much network traffic and data transfer http://www.bbc.co.uk handles every day) and it would be likely to break under the strain.

   The BBC have a remit to ensure that the service is available 99.9 percent of the time – and we tend to meet that target head-on – having a fixed IP address for every single redriect would almost certainly never be available 99.9 percent of the time.

8. OK, let’s go through this response bit by bit.

   “bbc.co.uk isn’t an official URL”. Well, taking the fact that it’s either a domain name or a URI aside, if it’s unofficial, then *stop your presenters from saying it on air*! “It can never have the same characteristics … as a full URL” – now that’s mush as well. I can merely cite what you’re accessing right now – neuro.me.uk works exactly as http://www.neuro.me.uk does, in terms of responding to browser requests for web pages. So that’s part one of your response refuted.

   Saying that bbc.co.uk can “only ever be a static IP address” is not factual, if all you’re suggesting is that it can only be a *single* address, as the root part of the domain can be pointed at several IP addresses, which are then used in a ’round-robin’ fashion. Do a lookup on something like microsoft.com – I just did, and it’s currently pointing to two IP addresses, amongst other things (MX and NS records). Nonetheless, the restriction that bbc.co.uk can’t be a CNAME alias, as http://www.bbc.co.uk is by pointing to http://www.bbc.net.uk, which itself has several A records pointing at it, does nothing to restrict it’s usage with regard to web hosting – *it’s current use illustrates that*!! So that’s part two of your response refuted.

   “If we started allowing that IP address to handle bbc.co.uk/anything …” Here’s a problem. Why set up a server to respond to bbc.co.uk in the first place? I remember a time when it simply displayed a page stating that it wasn’t a kosher site, and here were some links to other BBC sites (e.g. BBCi home, BBC One, Radio 1, Worldwide, etc). Now it redirects to http://www.bbc.co.uk. This implies it can be easily retooled to handle bbc.co.uk/something redirects.

   Let’s also look at the fact that presenters routinely dish out the URI as “bbc.co.uk forward slash something”, and people listening will put *that* into their browsers, and will not compensate for short-sightedness by the implementors at New Media or the policy makers in Commissioning, and put in “www” to the address box for you. I seriously doubt there would be a marked increase in traffic if the configuration was changed to cope, since the current usage on air must lead to a level of demand already, so therefore your “immense demand” remark makes no sense, as it must be coping with a level of demand as-is. Even if the demand increases, it’s straightforward, albeit not “trivial”, to place a load balancer in front of the server (or prospective servers) handling bbc.co.uk redirects. I’m sure New Media, amongst other BBC departments, must be investigating such technology after Mr. Dyke’s recent Creative Archive announcement. Alternately, change bbc.co.uk to use a server (or servers) in thdo or thny – it’s not like www-bogus.kw seems to do much else. So that’s part three of your response refuted.

   And, at the risk of getting even more up myself, I’ve got a very good idea of the amount of traffic the BBC farms handle – Telehouse Docklands peaks off at around 1 to 1.5 gigabits/sec, Telehouse NY at around 350 megabits/sec and San Jose and Kingswood Warren both peaking at about the 30-40 megabits/sec mark. If you’re talking just http://www.bbc.co.uk, www*.thdo and www*.thny all seem to hover around the 5-20 megabits/sec mark, with 2 and 3 digit per second hit counts. Given an average of about 3.5 megabits/sec transfer rate over the 36 or so www* machines, that’s about an average of 295 gigabytes a day per machine, so a fair estimate of http://www.bbc usage a day would be roughly 9.5 terabytes a day aggregated across Docklands and NY farms – noting that 3 of the 36 servers seemingly aren’t operational, and that it’s a guesstimate based on average throughput alone, and not considering peak/off-peak and ad-hoc usage, i.e. post-programme hits, one-off online chats, etc.

   Now, where were we … oh yes. “… having a fixed IP address for every single redriect [sic] would almost certainly never be available 99.9 percent of the time” This problem has such a trivial fix as to have next-to-no impact on current load and demand, and most certainly doesn’t require more IP addresses for each and every request placed upon the server already in place. Also, if you have a system in place which is believed not to comply with the service level remit of 99.9% (which by high availability standards isn’t that good – that’s about 8 and a half hours of downtime – most critical sites would consider that unacceptable), then why is there the insistence to let presenters refer to it on a regular basis? So that’s part four of your response refuted.

   … and in the end, I still remain wholly unconvinced and am baffled at any positive thought on how to correct this issue. I’m more than happy to demonstrate how trivial this probably is to work out, for a nominal fee of course :>

   The suggestion to make a trivial correction to a single server sitting in Kingswood Warren is unravelling a *very* interesting view on the BBC content network, and the unspoken attitudes by implementors and policy makers regarding this aberration of a URL. Keep em coming tho! You guys can either keep that big networked head in the sand, or you can take a bit of help when it’s offered 🙂

9. All of what your saying makes sense, but we’re not talking about ‘trivial fixes’ here – we’re talking about resilience on an absolutely huge network.

   Forget all that stuff about presenters – the times when a www is dropped are few and far between. We’ve researched this – and most people have half the brain to type in three consecutive w’s when they get redirected to the main BBC homepage (which was a policy and planning decision, by the way – to avoid giving people the impression that the who website was broken).

   Additionally, you obviously haven’t got a clue about deliverable terms of service – which is what I was talking about when I said that the BBC network must be available 99.9 percent of the time.

   In actual fact, the BBC network is available 100 percent of the time – but no-one in their right mind would attempt promise to deliver such availablility (not even on the most critical website), since all manner of unpredictable factors make it impossible to do so. The remaining 0.01 of a percent is to allow for artefacts.

   The simple fact is, all of the solutions you have offered (most of which are unworkable, if truth be told) have already been considered and rejected because of resilience that falls below acceptable levels.

   Also, your knowledge of the BBC network seems entirely limited to what you are able to access on the open www network – the fact alone that you seem to think that part of our server farm is based in Kingswood Warren belies this – our main server farm cluster is now based in Maidenhead (with others in THNY, THDL and San Hose) and has been for months – only a few bits of relatively unimportant kit are left in Kingswood.

   As for you being more than happy to demonstrate how trivial this probably is to work out (for a nominal fee, of course) – don’t call us, we’ll call you – once again, I’ve just waited in greybar land for what felt like an eternity whilst your website downloaded.

   Maybe it’s time to cough up a few quid for some decent web hosting instead of relying on a 256kbit/sec DSL line? (In plumbing circles, that’s known as trying to fit a nine foot turd down a three foot pipe).

   If you applied those standards to a BBC website, you’d be on the dole by lunchtime on your first day.

10. *cough* *splutter* I’ve put up with most of your rude comments so far, but these take the biscuit.

    I can’t forget the stuff about presenters, because presenters are *permitted* to drop the www by commissioning guidelines. Virtually every web and networking professional I’ve spoken to thinks the lack of a working redirect between the bbc.co.uk/somthing URLs and http://www.bbc.co.uk is ridiculous, and my solution is both elegant (takes seconds to implement on a sane Apache configuration) and just as resource light/heavy as the existing configuration running on bbc.co.uk right now. Again, my solution has been validated by others who I’ve spoken to about the situation. The mere fact that you say they are “unworkable” leads me to wonder to the rationale behind the rejection in the first place.

    Yes, my knowledge of the BBC network is limited mostly to what I know from publically accessible information. Given that the IP address bbc.co.uk is mapped to resolves back with a kw.bbc address, you’ll forgive my assumption that the box actually lives in KW. Oh, and also the fact that the the last hop on the route to that box has an IP address registered to KW, not Maidenhead. That really threw me off the scent.

    Again, in speaking to other people, my website loads acceptably to people in dialup, broadband and fast network lands, with times varying between 3 and 13 seconds, obviously varying with both their and my local network conditions – upstream usage, etc. I’ve accessed my own site from other peoples connections, be they dialup, ntl:, DSL, whatever, and i’ve never sat in “greybar land” for my page to load. May I respectfully suggest you find some new software? Or a new network? This high and mighty BBC network doesn’t seem to good when those browsing sites over domestic broadband can beat up download times over someone at the BBC 🙂

    I really don’t appreciate the “nine foot turd” comment, but my decision to host my own personal site on my own 1 meg ADSL (I quoted 256k because that’s the upstream, and thus the maximum speed that anyone can download anything from it) was purely a choice of convienence. I routinely host and manage sites in the US, Canada and on high speed networks here in the UK, including in Telehouse Docklands, not a few rooms away from the BBC servers. I cough up enough quid on hosting, thanks all the same, and given your comments over the last week or so, you’ll excuse me if I ignore any advice from yourself. Why are you so determined to attempt to show me up compared to your own – sorry, the BBC’s – shortcomings? I’m showing that a large UK corporation is neglecting a crucial part of it’s public network infrastructure, despite that same part being routinely referenced by it’s public-facing staff and contractors – and you’re saying that the BBC basically doesn’t give a damn. And you’re doing it in a particularly rude fashion. I believe I’ve been a bit cheeky to you, but at no point have I insulted you personally. Is this what I pay my licence fee for? Is this something that BBC employees are permitted to do to their “customers”?

    Anyway, in closing, I’m interested that you decided to continue this discussion on my website and not by email – got an identity to hide? Also, if I applied *your* standards of discussion to a BBC employee, I’d be looking for your books before the morning tea break.

    Oh, and at the risk of descending into pettiness, it’s spelled “San José”.

11. You need to chill out a bit my man – it’s only a bit of banter. If you don’t want skint knees, don’t go cycling without your kneepads.

    Where have I been rude to you?

    I have not insulted you personally in any way, but you’ve totally dodged the answers I gave you in the last post.

    Your knowledge of the BBC network *does* seem limited – from my perspective – and I’ve merely pointed that out to you.

    I’ll say it again – it’s all about resilience. Sure, some of your suggestions might be effective and reliable on another website – but not http://www.bbc.co.uk as far as I am aware.

    In answer to your other points:

    Presenters are NOT permitted to remove the www from a promotional URL – they are only premitted to drop it if the web address appears IN FULL on-screen. If and when presenters do drop the www in other circumstances, they do so of their own accord – which is a problem the BBC are trying to tackle.

    Secondly, you seem perfectly happy to fire off tirades about the BBCs website, but you’re a bit too sensitive about comments directed at your own personal website. But okay – I’ll try not to hurt your feelings any more.

    And, oh – where have I said the BBC “doesn’t give a damn”? No-where. That’s what you call mis-representing someone – and you’ve done it just because you’re angry.

    And I’m answering your points here, because you started a thread on an open forum. Had you emailed me, I would have replied pretty much in the same fashion.

    However, I’ll tell you what I’ll do: I’m attending the BBC Internet Operations meeting the week after next. I’ll get all of your points raised at the meeting and I’ll put forward your suggestions – and I’ll even ask for a report to be drafted.

    How does that sound?

12. Let’s see now, seeing that my obvious rampant rage is keeping me from making rational replies …

    You gave me one answer of note in your second last reply (the “research” reply) which didn’t really answer anything, and seemed to imply that by cleanly redirecting bbc.co.uk/something to http://www.bbc.co.uk/something without the user actually noticing, they’d think something was broken. I didn’t dodge anything, I ignored it.

    You have been rude to me, several times, whereas I’ve only been rude to the BBC in general, not to you individually. If I’m sensitive about comments regarding my own site, it’s because they appeared to be rather unfounded – as I’ve mentioned, others have no problems accessing my site in a timely manner. Considering I built my own server myself, and you probably had nothing to do with the main server farm builds, you’ll excuse me if my sensitivity level in a particular direction is high 🙂

    But anyway, to the points in hand.

    If they are permitted to drop part of the web address at all, then it is a legitimate truncation according to guidelines. Therefore, the network should be able to cope with that truncation in an effective manner, not dumbing down the response to the user, i.e. ‘oh, it’s still BBCi, that’s ok then’ and they have to figure out where the hell the content is that they were looking for. That extra click that they have to perform is kilobytes upon kilobytes of bandwidth used – and wasted – in loading the BBCi homepage, when they could have been redirected to the page they were actually wanting to see in the first place.

    As to the address being on screen, what if the address is being displayed in the lower quarter of the screen – a fair assumption, I think you’ll agree – and that the user is watching the programme with, let’s say for argument’s sake, a Sky digibox set to default (4:3 crop), and that they are watching this on their 28″ widescreen set with zoom-fill selected, thus cropping the upper and lower parts of the 4:3 already-cropped image, and removing or distorting the address. That means the only means they have of knowing the address is the presenter *legitimately* saying “go to bbc dot co …” Not an every-viewer-will-have-this-setup scenario, I admit, but a fair possibility I reckon 🙂

    Actually, a more realistic scenario is someone hearing the address, and just not seeing it on screen (looking away for some reason perhaps), or concentrating on the speech and not the caption.

    “… where have I said the BBC ‘doesn’t give a damn’?” – OK, I’m extrapolating that from what you’ve said previously, so I’ll give you that one. I didn’t misrepresent anything because I was angry though. I may have been incredulous at points, but never angry. Anger towards personas digital is invariably pointless, so I usually reserve that for people I actually know 🙂

    But OK, you can raise these points at an Ops meeting – that’s very cool. May I summarise my points thusly:

    – allowing presenters at any point to say “bbc dot co dot yoo kay forwardslash something” and not redirecting the url http://bbc.co.uk/something/ to http://www.bbc.co.uk/something/ is detrimental to the image of the BBC in terms of being able to operate an effective and usable Internet experience.

    – forcing the user to be redirected from http://bbc.co.uk/something to http://www.bbc.co.uk/ happens already, and the modification of a sane Apache config using the Redirect option will not increase the existing load on the box handling said requests at present by a measurable amount, since it is already using resources to – incorrectly – redirect such requests already.

    – forcing the user to be redirected from http://bbc.co.uk/something to http://www.bbc.co.uk/ is also a waste of the BBC’s bandwidth, as it requires a page load of the BBCi homepage, followed by (with luck) a page load of the URL the user was actually attempting to find. If the URL onscreen was a redirect to somewhere deeper in the tree, there may be several clicks, and thus page loads, required to find the correct page. This increases the load on the BBC’s servers and network unnecessarily.

    I think that covers the crux and kernel of it all. I dunno if you’ll believe me, as I *have* been rather cheeky at points 🙂 but I do appreciate the response of a BBC employee to my meagre grumble (bah, humbug, grumble, raaar), even on a Sunday when I was trying to get ready to watch the Dundee United game – on the BBC no less! 🙂

13. This is a fascinating discussion. I’m pretty sure that Radio 4 presenters say “bbc.co.uk/” more often than not. I’m going to pay special attention over the next few days in case it’s just my selective hearing glossing over the “www”.

    I thought I’d look for networks that are probably of a similar size as the BBCs and came up with the US Government and note that they can manage to get this right for http://whitehouse.gov / http://www.whitehouse.gov and also for their state department at http://state.gov / http://www.state.gov. Presumably every other one of their sites gets this right, but I couldn’t be bothered to check. I also see that Microsoft can get this right too on microsoft.com and they run lots of xxx.microsoft.com sites in addition to their main http://www.microsoft.com.

    I am amazed that setting the system up to redirect web browser requests for http://bbc.co.uk/xxx to http://www.bbc.co.uk/xxx is so difficult when other large organisations seem to be able to do it…

14. I made an interesting discovery a few minutes ago … I was writing a blog entry about how tabloidy Horizon has become (I’ll post it later) and was googling for the recent documentaries on air safety (“Crowded Skies”), when I stumbled across a search result from the BBC’s site with the url http://bbc.net.uk/ …

    Try clicking on that – *it works*. http://bbc.net.uk/scotland/ … http://bbc.net.uk/bbcone/ … and so on. When I looked at the host information, the following two scraps of information were extremely interesting:

    intrepid:~> host -t any bbc.net.uk
    bbc.net.uk A 212.58.240.120
    bbc.net.uk TXT “BBC Intelligent Load Balancing Domain”

    Sidenote: If you do that for yourself, you’ll also notice in the SOA record, which I omitted, includes the contact address of bofh at bbc.co.uk 🙂

    Now, I know that http://www.bbc.co.uk aliases to http://www.bbc.net.uk which has several IP addresses connected to it by A records, but here is a note saying that it’s a load balancing domain. More interesting information.

    So basically, the A record for bbc.co.uk could be changed to the A record(s) for bbc.net.uk, which points to the BBC web farm, and would make a lot of the problems I’ve documented here go away. Very rapidly. I’d be really interested to hear the operational reason why bbc.co.uk points in a different direction from bbc.net.uk … 🙂

15. just thought I’d stick a comment in to keep everyone reading up-to-date … I’ve been chatting to Martin by email — a much more professional way to do business 🙂 — and the bottom line is that he brought this topic up at the Online Operations meeting in Maidenhead, and they all had a gab about it. The outcome was that they agreed to investigate it again, as apparently it’s been looked at before, and they’ll talk about it again at their next meeting in October.

    So hopefully this will get sorted, and not just “acht, who cares”-brushed-under-the-carpet sorted.

    Thanks to Martin for getting back to me 🙂

16. Well, another reply from a BBC technical bod via Martin, and it’s rather unsatisfactory (surprise). The basic view seems to be “we shouldn’t be using bbc.co.uk in a URL, as it looks unprofessional, because we use all these protocols like real and chat and suchlike, so we don’t want to use bbc.co.uk just for web – oh, and we don’t mind offloading sense to the users, who we notice don’t care about getting the wrong page, and then manually feeding in www themselves – and it doesn’t happen often enough for us to care anyway” (I paraphrase, of course).

    I’m getting more than irritated about this now – announcers continually say “bbc.co.uk” instead of “www.bbc.co.uk” as the technical guys would clearly rather (and lazily) they say instead. It’s not just announcers using it while the full http://www... is onscreen, it’s all the bloody time! I’ve even seen bbc.co.uk used in a caption on BBC News 24 – http://neuro.me.uk/blogsupport/img/bbcdotcodotuk.clickonline.png !

    Meanwhile, http://bbc.co.uk/ appears to *not actually work anymore*, thus making every instance of a presenter using bbc.co.uk seem that much more unprofessional! Horrible!

17. Incidentally, Radio4’s own meeja commentary programme, The Message regularly announces its URL as http://bbc.co.uk/radio4/factual/themessage rather than the actual URL of http://www.bbc.co.uk/radio4/factual/themessage.shtml . At least with the addition of a www, they have the decency to redirect you to the right page…

18. Ah, possible confusion between me of the post above and Martin of the BBC. I am not he, and don’t even play him in panto.

19. Well, I’ve had an update from Martin, nothing I can really say right now, but there will hopefully be an update soon.

20. Wow! I’ve just trawled through all this, and I must say I’m impressed by your perseverance. I’m not at all surprised by the BBC line, though I am surprised that one of their (our!) staff would be quite so rude to a licence-payer in an open forum. I’ve come up against this myself many a time: you try to help a big corporation out by pointing out a usability flaw in their site or a bug in their software, and all you get for your trouble is condescension at best.

    As for users being smart enough to type in the “www” themselves, I take it they’ve never done any real-world user-testing. The average user is your mum, or your grandad, and if you don’t spell everything out for them they realy do have trouble. It isn’t being condescending to cater for people’s needs, quite the opposite. If a commercial company want to steer customers away form their site then that’s fine, but the BBC are there to serve the public, and that means all the public.

    This Martin guy must have spent longer talking to you than it would have taken to fix by now. Sounds like something positive might come of it, though: any news? If you don’t get any joy, how about contacting Watchdog? They seem to like shaming companies into taking action in the interests of their customers. Sorry, I mean employers.

    PS do I take it that our licence fee is being used to pay for all these stateside server farms catering for transatlantic traffic?

21. this issue has perplexed me for some time too – respect to neuro for trying to get something done.

    as you mention above there’s an incredibly easy solution to this that will not overload a server …
    run a very very simple webserver on bbc.co.uk with the following line in its config:
    Redirect / http://www.bbc.co.uk/
    they seem to be redirecting bbc.co.uk/anything to http://www.bbc.co.uk (losing the /anything), maybe with a line like:
    Redirect /* http://www.bbc.co.uk/
    which _must_ be causing as much load on the server as a correctly functioning redirector would.

    and even if users _can_ get round this problem, ?the UK’s most visited website? could save the whole world a load of time and stress by redirecting requests to the desired page rather than their homepage.

    the beeb say that their services (or any fix to this issue) must be “available 99.9 percent of the time” – yet currently bbc.co.uk/thepageyouwant is available ZERO percent of the time.

    this problem makes the bbc’s presenters appear foolish when if they give out a url without the ‘www’. perhaps the bigger bosses of the beeb like to make fools of their presenters and customers. the whole situation seems typical of the image i get from bbc employees – the organisation can make mountains out of irrelevant things and can fail to see the obvious no matter how obvious, or how many times it is pointed out.

    oh well, at least we’ve got a www to rant on.

22. The (www.)bbc.net.uk A record is served by “BBC Intelligent Load Balancing Domain”, bbc.co.uk is served by BIND – you can’t return a CNAME of http://www.bbc.net.uk for a request on bbc.co.uk. http://www.bbc.co.uk returns a CNAME to http://www.bbc.net.uk, which then can work out where you are on the Internet and return the best location for you to be served. It can either return an ip within that location, (a h/ware single IP loadbalancer address) or randomly returned addresses (it’s quick to remove/weight IPs returned in the dns lb config). For significant money and time (don’t know if this even existed 7 years ago) you can setup global routing solutions that mean your single IP address can go to different locations,however we’ve saved licence fee payers money with a quick solution – I only heard about such a thing recently. We could have but didn’t hack around with BIND code to have bbc.co.uk return an intellegently returned single IP A record – this would have been a pain for many reasons. The Intellegent DNS servers
    aren’t that smart, so can’t do bbc.co.uk work such as MX records or other domain stuff for the other uses. A BIND based round robin DNS response still leaves dead boxes returned randomly (default DNS timeout is hours) or more importantly dead network h/ware LB solutions. This wouldn’t meet 99.9% availability and I can confirm we see downtime with many other sites. A very fast single box/LB as the A record for bbc.co.uk could probably deal with the load, existing ones didn’t mind, and if we’re lucky it stay up 100% of the time by pure chance. Changing the rules to insist people start saying “WWW.bbc.co.uk” would be the better solution IMO!

23. bbc.co.uk as opposed to http://www.bbc.co.uk is now apparently ingrained in presenters minds, and it takes prodding and poking to make them add www. to the start.

    The BBC could save licence fee money by having a single television channel broadcasting in 425-line PAL in black and white with mono audio, but that excuse doesn’t merit refusal to evolve technology in favour of saving money – this is why we have such things as 625-line PAL, NICAM, stereo FM, widescreen, MPEG2-DVB, and so on.

    Changing the rules doesn’t solve the problem.

24. May 1

    Put the date in your calendar.

    All your troubles will be over!

    The “www” in a BBC URL will be gone!

    Redirects will work!

    On-air promos and trails will not have “www” in them!

    Presenters won’t need to read out “www” any more!

25. Got a source for that information?

26. As I gaze into my crystal ball. . . . The colour blue comes rushing up at me. . . and a date. . . May the 1st. . (unless someone important karks it). . .

27. More fun from the “What’s Wrong with the BBC Website” Wiki – available at http://cheerleader.yoz.com/w/WhatsWrongWithTheBbcWebsite – looks like an (“The”?) insider at el beeb is getting stroppy.

    “You seems like a smart person, why not just type the “www”? […] More crap. Just type the www for God’s sake – are you incapable of doing that?. The pages will work fine if you do – it is the *W*orld *W*ide *W*eb after all. Who cares anyway? It’ll be sorted by May 1 anyway. Argument over.”

    Obligating the *visitors* to correct technical mistakes made by the BBC is disgusting behavior to exhibit, and I hope that the attitude of this “insider” is not consistent with others in the Corporation. I also hope it’s not the same “insider” who posted above.

    Also, saying that the argument is “over” simply because they plan to fix something is not good enough. The argument will be over when there is an explanation as to *why* there should have been an argument in the bloody first place.

    I’ve been kind of schtum over the planned 1st May “troubles will be over” fix – I will probably break silence at some point near or at the time.

28. Oi!, I posted above, an no, someone else posted the ‘users can just type www’. Lets just take a breather for a moment here. Most people who write and blog about this sort of thing are pretty tech savy / alpha geeks. I’ve witnessed user testing, (yes Steve who posted above, we do do it) and not just 20 something white Londoners, but middle aged regional house wifes, you should see how they actually use the internet, or even computers!

    I think this arguement is bigger that “..apache servers config re-route blah blah..” hopefully it’s about trying to offer users the best service we can, and. . . wait for it. . talking to people like yourself and passing that feedback on.

    PS, I’m not a techie so please don’t come back at me with code talk.

    PPS. http://bbc.co.uk works fine, as does http://www.bbc.co.uk and http://www.bbc.com and bbc.com isn’t that enough.

29. Steve, the license fee doesn’t pay for transatlantic traffic. The international part of the infrastructure is paid for by the World Service, which is funded directly by the Foreign Office, not through the license fee.

    A part of the infrastructure in New York is used as a backup in case anything serious goes wrong in London, and that *is* paid for by the license fee.

30. If, and you should be before ranting, anyone’s interested in how the licence fee is used. point yon browser(s) to http://www.bbc.co.uk/info/licencefee/

    Here’s my strap line: The BBC, you love it really.

31. eyedropper: “http://bbc.co.uk works fine, as does http://www.bbc.co.uk and http://www.bbc.com and bbc.com isn’t that enough”

    No, it’s not enough. A presenter says “please visit our site at bbc.co.uk/somethingorother”, and when Joe Average Licence Fee Payer keys that into their web browser, they get the BBCi front page. Now they have to drill through something like one to four layers of navigation and page loads to find what they were looking for. This is typical of the kind of attitude I’ve been railing against for months, e.g. “it works fine”, “we know better”, “they can just put the www in themselves” … even with a solution on the horizon, there’s still this biting the hand that feeds …

    I’m ranting *because* I’m a licence fee payer, and if it’s *my* money that’s being spent on a badly designed system, then by $deity I’ll have my bloody say about it.

    Also, eyedropper, I wasn’t attempting to name you specifically, just whoever this “insider” was who was posting away on the wiki. Use a similar nick to attempt anonymity and prepare to be pointed at 🙂

32. My point about the redirects with/without www related just to the homepage. as in, “hmm, what’s the bbc’s website?” user can type in a number of options.

    As regarding presenters reading out URLs, and because I’m a caring chap who’s actually trying to help out, I rang our Editorial Policy dept, and also the Senior Announcer at Radio 4, they confirmed to me that (on Radio 4 at least) they ALWAYS read out the FULL address: w w w dot bee bee see dot co dot uk forward slash, radio four . Sometimes if they’re short on time, they’ll say ‘go to the radio four website’ .

    I hope this draws the issue to a close, yes, perhaps on some TV broadcasts sometimes www is omited,(probably live ones are the worst culprits) but you’ve not much evidence of how many people this actually affects. As part of my job I also look through the customer feed back logs, which contain mails sent to the BBC general email query address. To my mind, I can’t recall ever receiving a complaint of this nature.

    Thirdly both myself and others have tried to answer your questions to the best of our abilities. Fourthly, It’s probably cost the BBC in people and time and infrastructure the cost of your license fee to find all this out and right back to you! (that last bit’s in jest btw)

    Finally, I’d like to ask this. Do you feel you get value for money? Your license fee breaks down to 33p a day, do you get a good service for that?

33. > you’ve not much evidence of how many people
    > this actually affects

    Of course I don’t, I don’t work anywhere near the BBC, hence have no access to any data. I’d suggest web server logs on the machine(s) doing the redirects from bbc.co.uk to http://www... would be more pertinent than customer feedback logs.

    > both myself and others have tried to answer
    > your questions to the best of our abilities

    Is that an excuse or a fob-off?

    > Do you feel you get value for money?

    Not if it takes over 8 months to get something done which could have and should have taken much less time to achieve, with much less backbiting from those within the BBC. I think Martin eventually realised what I was talking about after much banging on the head – if you reread the whole of this web page, I hope you will too.

34. Right, the ‘bloody stupid’ comment on my blog refers to the fact that the discussion has gone on for nearly a year.

    In that time you’ve not actually collected how many times a day web address are read out on air. so you’ve still NO data, only a theory.

    So despite finding out for you that Radio 4 (amongst others) DO read out WWW which counters your theory I’ve rang our Internet Ops team for some ball park figures. A rough guesstimate of requests across the BBC network is 5000+ a second. so that’s 300,000+ a minute, minimum. The redirect server (http://bogons.bbc.co.uk/) that catches ‘bogus log ons’ (hence the name) gets 2 hits per minute. This also includes users typos etc.

    Lets review those numbers shall we 300,000 vs 2 a minute.

    At worst, the user arrives at the homepage (not a 404), where there’s a huge and helpful search box and a categories list. AND it’s being fully address from the start of May.

    I also think your analogy of ‘reported sexual attacks’ is out of line on something as trivial as a discussion on server redirects.

    I tried to be helpful, with out so much as a thanks, lets just move on eh? It’s spring, bit of sunshine, walk in the park. . .

35. I did state that my analogy was merely to illustrate that things happen in the world that go unreported, it was the only thing I could think of from the top of my head that was “real world” and documented.

    I know that some presenters read out the www part of the address, I didn’t need someone to find that out for me, but I’ve heard it omitted often enough to get annoyed about it. Others agree with me. What does it matter that I have no hard evidence from the servers or the feedback logs? I *hear* enough evidence every time a presenter omits www and I find that the address *as read* doesn’t bloody work. Imagine a presenter on Swap Shop decided that there was no point adding “01” to the 01 811 8055 phone number because the viewers could figure it out for themselves?

    Incidentally, 2 hits a minute equates to 2,880 a day – that’s just over a million a year. If each of those hits is a person following a redirect, that’s a million people finding out the hard way that the BBC don’t actually care about usability. Even if it’s a fraction of that, that’s still a sizable number of people having to work harder to visit the site. Do you thing someone with a screen reader has half an hour to spare to listen to the BBCi main page read out to them, not to mention the layers they have to click through to find out where they were actually trying to get to in the first place?

    Still unconvinced of your argument. The homepage might as well be a 404, as it’s *not the page they wanted to see when typing in the address*. I’ll say thanks when I actually see some courtesy.

36. eyedropper – you state two things above:

    1. Right, the ‘bloody stupid’ comment on my blog refers to the fact that the discussion has gone on for nearly a year.

    2. In that time you’ve not actually collected how many times a day web address are read out on air. so you’ve still NO data, only a theory.

    re. 1.
    is the comment above rhetorical ? is it neuro’s (or the license payers’) fault that this has gone on over a year ?

    re. 2.
    you expect the customers to fix your problems ? or tell you all the ins and outs of your own business ?
    when a company has this attitude i usually just stop dealing with them. not really an option here since the BBC is a unique company. and since we pay for it i think we have a right to ask for decent service.

    of the 2,880 redirects a day i expect over 1,000 of these to have lead users to swear at the screen. in my experience “middle aged regional house wifes” get frustrated easily enough when using the internet. i think it is such a shame that britain’s most respected website frustrates these timid people. the bbc’s decisions and actions (to read out incomplete web addresses and then to not redirect web requests properly) is actually very damaging to the confidence and learning process of these “middle aged regional house wifes”.

    i like the strapline ‘The BBC, you love it really.’
    it is a true statement, there’s a lot i love about the bbc, and as a whole i probably love it (and it’s content) more than i get frustrated by it (and it’s practices).
    however, the _real_ reason i like your strapline is because you included the word ‘really’. how apt.
    if the bbc didn’t annoy me with issues like this then i could say:
    ‘The BBC, don’t you just love it.’

    sadly i don’t _just_ love it, i get very frustrated with it too.

    bring on may the 1st …
    (just for all the “middle aged regional house wifes”, obviously us techies have coped just fine all along *hmmm*)

    ps.
    my two fave quotes on this subject:

    “… the boffins at BBC Technology, famous for having almost limitless technical skill, infamous for having similarly huge egos and stubbornness.”
    (suitable for bbc managers too ?)

    “… that’s a million people finding out the hard way that the BBC don’t actually care about usability.”
    (sums up the feeling we all get from this situation – do the BBC care about us ?)

37. For F**ks sake!! you people really need to get a life! go to the pub or go travelling or something!
    Jesus!!!!!!!!!!

38. Mr Moon, using a non sequitur like “get a life” when talking to people who clearly do ‘have lives’ is pretty insulting. It is possible to have a regular working and social life, and still get pissed off about this kinda stuff.

39. Friday’s The Message (Radio 4’s flagship media program) managed to read their URL out as “bbc.co.uk/radio4/factual/themessage” not once but twice, as did the continuity announcer straight afterwards.

    The continuity announcer clearly didn’t think that listeners were able to intelligently amend the URL as she explicitly specified forward slashes.

    Incidentally, even adding the www. doesn’t result in the *correct* URL, which is http://www.bbc.co.uk/radio4/factual/themessage.shtml, but at least you get an intelligent redirect.

40. They’re not even supposed to do that – only top level URLs are supposed to be given out according to guidelines.

41. *Une grande victoire!*

    http://bbc.co.uk/bbcone/ … http://bbc.co.uk/radio1/ … http://bbc.co.uk/health/ … http://bbc.co.uk/news/ … All working at last! Thanks to everyone who commented here and emailed me in support, to others who noticed that something stupid could be fixed, and to Martin at the BBC for championing my cause!

    And there’s more to come …

42. Kudos to you!

43. The question is, was it really worth the thousands of pounds of license fee payers money that it took to fix this? No, really, there is about 40k’s worth or hardware, not to mention the installation, setup and admin time taken to fix this. Was that really the best way to spend the license fee? O, and I forgot to mention the ongoing costs to the organisations that would be required to support and maintain something like this. Perhaps you should think about the true costs of thing before you start your little ‘campaigns’.

44. I suggested a single line change to an Apache configuration to fix this problem, the BBC has chosen to deploy £40K of equipment (according to you, whoever you are). That’s their perogative, but I cannot and will not feel guilty about the fact that this configuration should have been running from the outset. The instant a presenter omitted “www” should have been the point where someone at New Media (or wherever) should have said “well, if one person does it, more might – let’s get tooled up for this”. Either that, or a much more stringent policy on reading out URLs over the air should have been put in place. The Beeb decided against the latter, so where else was there to go?

    The level of arrogance I’ve seen in responses from BBC employees has been absolutely astounding. Services are deployed with shortcomings, and when those shortcomings are identified the natural response seems to be to attack those who report it, not to respond with humility and grace.

    I appear to have been instrumental in this change coming about – would I do it all again? Almost certainly. If I see something similar, will I make just as loud a noise about it? You betcha.

    Perhaps the BBC should consider the *proper* way to deploy a new media service before having to bolt on fixes later on. Nonetheless, I really can’t get worked up about the BBC having to spend £40K on something when their revenue from the licence fee in 2002/2003 was a bawhair off £2.66bn.

45. Sorry. Marketters are FAR more influencial than you…

46. I have some information to refute that, but sadly I can’t publish it 🙂

47. Well, the BBC website is now bbc.co.uk, not BBCi! Woo! That’s the last change I was promised would happen. Over the weekend – exactly on the 1st of May in fact – new style promo presentation started on BBC One, where the URL has “www.” omitted on-screen. Now the BBCi brand has been dropped from the website in favour of bbc.co.uk.

    Just goes to show what can happen if you are insistent and polite enough to the Beeb. 🙂 Thanks again to everyone who supported and championed my cause, both inside and outside the BBC.

48. So the entire of the BBCi rebrand had absolutely NOTHING to do with a brand that didn’t work in that context and everything to do with you ranting about incorrect URLs?

    Sorry whilst I roll around on the floor laughing.

    Oh and for the record, I WORKED on the rebrand.

49. If you’ve no idea what went on in the months I was silent between comments on this page, then you’re probably not in full possession of the facts 🙂 I’d reckon that (excuse me while I extrapolate here) your PHB said “media person, rebrand our site thusly, as laid down by the Mighty Marketing Droid Army, and don’t ask any more questions before I beat you with a dogeared copy of the Hutton Report” and not “rebrand our site because [complete history of rebrand rationale from minute one]”.

    It was my constant nagging that got the topic raised internally, and I therefore consider this discussion influential, if not instrumental, to the rebrand taking place. It’s not what I was after, but it’s a nice bonus on top of the redirects being enabled properly.

    If you can prove otherwise, anonymous supposed employee of the BBC (which at least your posting IP address seems to confirm), please do so.

50. er..no

    you’re kidding yourself if you think that the marketing people at the BBC(i/co.uk) ever look at the Web, never mind your site

    kudos to your interesting (and entirely sensible) campaign, but here’s the real deal

    1) (2 years ago) idiots in marketing can’t get their head around the fact that the Net and interactive teevee are different
    2) idiots think the best way to promote two separate platforms is to merge them into one (like telling people to tune to 97-99FM to watch BBC One)
    3) BBCi introduced over everyone’s objections
    4) 2 years on, the forces of the sensible triumph (due to lots of things, of which, believe me, this site isnt one, can anyone say ‘upcoming Top40’?)
    5) as a result of this ‘rebranding’, work is done on the way that the bbc balances its dns requests
    6) bingo – finally bbc.co.uk/foo works

    good stuff though – are you the same neuro that work(s/ed) on smoothie?

Comments are closed.