davew’s journal

HEAnet is hiring – there’s an advert in the Irish Times and on our site. The positions are Experienced Systems Administrator, Senior Projects Manager and Senior Network Engineer.

So the Department of Communications, Marine and Natural resources puts out a press release about the transfer of responsibility of the IE Domain to ComReg. It’s a slow news day, and it ends up on Five Seven Live and with a string of ever more sensational website headlines. ComReg to take over .ie domain name. ComReg to take over control of domain name. Government to take control of dot-ie. Irish gov to sieze control of .ie.

Here’s a game. One of these headlines is from the Register. Guess which one?

So what’s the story? Well, section 31 of the E-Commerce Act, 2000 gives the Minister the authority to regulate the IE domain. It appears that this will now be done, and that the regulator will be ComReg. I’ve yet to see any evidence that IEDR Ltd. will be shut down or that operational responsibility is going to go anywhere. What I do see is that IEDR Ltd. will now be accountable to ComReg, and so accountable to the rest of us.

I’m both pleased and sorry to see this, though I aim no displeasure at the Government or ComReg. The only thing I wish is that this was not necessary in the first place. Few people will dispute that the IEDR needs more accountability, but it was incumbent in the industry as a whole to ensure that that accountability was there, or to provide a credible alternative. When the first slipped through our fingers, we failed to provide the second, and so we look to Government to provide it for us.

So, this I think is largely a good thing, as it gives us a badly needed venue for input into how the IEDR is run. Now the responsibility returns to us again. The problem with the IEDR is not the rules. We came up with the rules. Remember the One Domain Per Holder Working group? That was representatives from industry, answerable to anyone who was interested enough to participate in the forum, and crafted the rules we still operate under. The rules are not the problem. The prices are not the problem. The problem is – cliché time – the lack of an open and transparent process to change the rules.

Now, after letting this go once, ComReg can give us the opportunity to recreate it – but only if we tell them what we need. It’s incumbent upon us, as users and operators, to come up with a workable system and propose it to the regulator. Until I see this happen, I don’t wanna hear anything about brown paper bags. We are now getting exactly we asked for.

A friend was asking me about my opinion on the rfc-ignorant.org blacklist.

So my question is this: Anal RFC adherence & dropping mail from those who don’t – good, bad or ugly ?

While I think that every network should have the right to choose what traffic it does and does not wish to listen to, I think that blacklists should be a last resort. Anything which reduces reliable connectivity reduces the stability of the internet and so there should be a strong case that the internet is made a better place by blocking them.
Read the rest of this entry »

ICANN has today insisted that VeriSign suspend the SiteFinder service, and restore the .com and .net top-level domains to the way they were operated prior to 15 September 2003. If VeriSign does not comply with this demand by 6:00 PM PDT on 4 October 2003, ICANN will be forced to take the steps necessary to enforce VeriSign’s contractual obligations.

This is fascinating stuff. ICANN was born running against the clock of the original contract with Network Solutions. Just how much control Netsol->Verisign can exercise on the root and COM/NET zones has never truly been observed – noone’s ever had to push quite this far before. The next few weeks will be interesting, and there is no guarantee that the internet will be more stable by the end of it.

I guess this makes it official. HEAnet launches Ireland’s first production IPv6 service.

I saw a story on Slashdot about the creation of the Mozilla Foundation as a new legal home for mozilla.org, and worried. Immediately afterwards I saw a story that said AOL had fired the remaining Mozilla developers. Okay, so the number of volunteer developers has outgrown the number of AOL employeers working on Mozilla, and a bunch of the developers have already been snapped up by other companies. But I always felt that AOL’s support was a safety blanket and, say what you like about the company, vital to keep the Mozilla wheels turning. Is anyone else worried?

More from our MEPs. Mary Banotti’s assistant replied with

We have received your email in the office and will bring to the attention of Mary Banotti.

From Niall Andrews’ office, I received this note which makes me wonder – I thought that his party were in support of the legislation. I’ll follow this with a phone call in a few days.

On behalf of Mr. Niall Andrews, MEP, I wish to acknowledge receipt of your email and to let you know that you have Niall’s support in this matter

Pat Cox’s office, who already sent me an explanation of his stance thus far, responded to my further questions.

As the committee will continue to meet between now and September I cannot answer your questions at this point. However I have forwarded you email to the Liberal Group’s representative and they have noted your concerns.

You can track the proposal through the Europa Legalisative Observatory. http://wwwdb.europarl.eu.int/dors/oeil/en/default.htm

Of course, I expect they all get an officeful of emails every day. I’m pleased with the clueful and reasonable reponse from Mr. Cox’s office, but I’ll follow up all three with phone calls and see if I can present myself as a clueful person to someone appropriate.

The Advertising Standards Authority of Ireland has upheld a number of complaints about Eircom’s advertising of the i-stream product. There’s a brief note in RTE News about it, but the complaints in full, Eircom’s responses, and the conclusions of the ASAI are detailed in their report.

At 14:52, I sent the previous message to Pat Cox, having already sent it to Niall Andrews and Mary Banotti. At 14:58, this message arrived from his assistant:
Read the rest of this entry »

Draft letter to my MEPs. I don’t know if this will get sent as is – too late to email, I think, and I don’t have access to a non-work fax. However, it crystallises my thoughts in case I find myself on the phone to someone useful. Feel free to borrow from this or suggest improvements/corrections in the comments; if you are contacting your representatives, though, I’d prefer if you rework this rather than send it verbatim.
Read the rest of this entry »

Software patents in Europe are a done deal according to Electric News. More soon, but the vote appears to be happening on Monday. FF and FG are in favour, Lab and Greens against – Dubliners might like to contract Mr. Niall Andrews, MEP and Mrs Mary Banotti, MEP.

Damn, the (more…) links aren’t working right here. Give me a bit while I find out what’s wrong. In the meantime, here’s a copy of the press release.

I’m quoted in a press release!

Irish Internet Neutral Exchange Implements Native IPv6

(31.05.2003) INEX, the Dublin-based Internet exchange, announced today the implementation of native IPv6 (Internet Protocol version 6) switching facilities at its peering centre in Clondalkin.

Read the rest of this entry »

Mozilla spamfilters are now catching everything spamassassin is catching, and then some. It’s still hitting me with some false positives, though, so it’ll be a while before I let it divert stuff into the “junk” folder. I don’t maintain an address book, so I’ve got no whitelist in Mozilla; this is raising my false-positive rate a bit. It’s heartening to see the little trash icon turning up though. This shit’s good. And it’s very easy to teach it; you hit the junk button just before you hit delete (next to it), or you untick the box if it’s a false positive.

It’s coping okay with IMAP, although life does get interesting if I open a folder with a lot of new messages (like active mailing lists I’ve not been paying much attention to – I’m a procmail bunny). It seems to download each mail in sequence so it can mark them, and this happens sort of in the background. With a stupid number of new mails and some stupid behaviour on the part of the user (um, me), I can confuse things somewhat, but I’ve not succeeded in doing any damage that couldn’t be undone by leaving the folder and going back in.

Favourite missing feature right now: find a way to store the spamfilter’s state online so I can share spamfilter intelligence between my work machine and my laptop.

Shameless promotion here. Prof. James Watson is giving a lecture in TCD this evening at 18:30 UTC+1, called “50 Years of DNA”, and it’s being streamed live. Test streams are up now so you can get the right software downloaded – windows media only, I’m afraid.

Worldcom is now MCI again, just when I’d finally gotten used to calling them by the W name.

This is my little bit of the Mandelbrot set. I got it from the mandelbit generator. You can get one there too.

Back from holidays, back from easter. I ran a dist-upgrade on my debian laptop at the weekend, and got Mozilla 1.3 with its Bayseyaiaiain spam filters. The build seemed to work pretty well, so I’m trying it in work. The filters need training; I’ve fed it the contents of my spamassassin-marked folder, which seems to lead to a fair number of false positives initially (user tells me that all mail is spam! holy shit!). I’ll report how it copes as time goes on.

Ok, so BitTorrent got my attention. Here’s Red Hat with the bright idea of releasing Red Hat 9 a week early to RHN subscribers, and BitTorrent goes and blows that idea out of the water.

Keeping an open mind, it’s a very interesting approach to the problem of mass downloading. It seems to be based on the assumption that if people are downloading, they have some unused bandwidth they can upload with. It splits the files into small blocks a bit like rsync. If the upload-bandwidth assumption is true, and if the blocks are small enough and the file large enough, it stands to reason that everyone should be able to attain wire speed.

Of course, there are a bunch of unknowns about this. Is grabbing blocks from random sources really a good way to do this sort of thing? You do lose control over the source, for example, so if you have high-speed access to a local mirror, you might lose out (unless you’re lucky enough to find a source that’s close to the shops, network wise). If everyone starts downloading from 30 simultaneous sources instead of one, what’s that going to do to things like CEF tables in core routers?

But it’s a neat idea, there’s no denying that. I’d love to be able to set up a “mirror” so that if a customer of mine wanted a (legitimate) copy of Red Hat, they wouldn’t need to know about my site at all; it would just default to being the handiest place to download blocks from. I don’t think the algorithm as it stands permits such a thing, sadly.