Theo's blog

When we first found our Jordan was pregnant last year, we decided we needed to move to a larger house. After spending a while looking around, we came to the conlusion we were going to have to move out of Docklands if we wanted anything reasonably sized, as it seemed there was absolutly nothing larger than 3 bedrooms that were half decent.

Just as we'd found a lovely 5 bedroom house in Oxford, I received a phone call from a letting agent; "I've just had a 6 bedroom house come on the market 300 metres away from your current place, are you still looking?"

What the hell I thought, nothing to loose (other than the extortionate rental costs). So we plodded over and had a look.

The house was perfect for us; 4 storey, 6 bedrooms, recently rennovated, wooden floors nearly everwhere (great for messy kids!), and best of all for me, a massive room that made a perfect office, overlooking the Millenium Dome (now the O2 Dome) and the river. Couldn't ask for a better view, short of renting out the top floor of Canary Wharf! It even had a large balcony the size of the current flat, and a great big (for London) patio undeneath .. so some space for Demi to have a paddling pool! Being only a few hundred metres away from the old hosue it meant that Demi would not need to go to a different nursery, and I would not need to worry about having 50mbit/sec+ internet, as I still had line of sight to two of the UK's main datacentres.

Within a few minutes we agreed on a price and decided to take the place.

That was 7 months ago.

Sounds great, right?

Right.

However, One of it's major charecteristics is also one of it's major downsides: Being on the Thames.

Firstly, in the summer, on a very hot day, the smell can get pretty, well, strange ... a bit like yeast. It's not really a bad smell: i kind of like it. It reminds me of the French canals when I was very young. However, Jordan really doesn't seem to like it.

Secondly, and this one is a lot worse than the smell: THE BLOODY PARTY BOATS. Yes, you know - the booze cruises that take sober people on board, and go up and down the river playing really loud music, getting them completely wasted. I'm sure it's great fun for the people on the boats (i've done one once, and it was fun, yes). But we have a 1 month old and 3 year old trying to bloody sleep at midnight. Argh. Of course, we didn't know this before, as although we'd been living in the area for a long time, we'd been living on the Canary Wharf docks instead. Because the hosue is Grade II listed, there is no double glazing: so the neighbours have no problems, just us.

Now, if someone wanted to open a nightclub 50 metres away from someones house, it's highly unlikely they'd be given permission by the council... But, who gives the boats permission to park an (open topped) nightclub outside my house 15 or 20 times a night?

Moral of the story: Don't live on the riverside unless you have double glazing or like lots of noise in the summer.

I got pissed off with WordPress, so i tried TextPattern. I got pissed off with that, so tried MovableType. None of them did what i wanted - a simple, plain text editor interface without any fancy javascript editors or such.

Mosly bizzarely, none of them appeared to output XHTML as the content within the RSS (although maybe they can, but I wasn't in the mood for hacking someone elses PHP, *shiver*), just encoded HTML - which feels just so wrong in 2006.

Note: Now i see what Pookey was ranting on about when whining about the problems with writting a real world RSS parser for Feed Banana. (Sorry i can't 'pingback/traceback/ifconfigback' you pookey - i've not worked it out yet)

So, 10 minutes later, an XSLT template addition to the current site, a few simple bash scripts, with moderate sprinklings of XMLStar, and i'm now running a fully web 0.1 compliant blogging "thing" (whatever the in name for them is these days, god knows - iBlogs maybe? :)).

Anyhow, all that is left to do by the looks of it is add traceback/pingback/whatever the hell it's called, and i'll be right up with the Web 0.2 guys. Ohh, and an RSS or Atom feed - once i work out the difference.

My how the web has changed in 3 years. I leave web stuff to go into programming real applications, and look how it all changes. I feel like an old man.

PS: Who's Scooble? :-P

Update: Upon further investigation of this whole RSS thing, i've discovered:

RSS is an XML-based format, which looks much like related formats such as HTML. Rather than including presentation information, though, the elements available in an RSS file are strictly limited to those required to define such things as a headline, a brief blurb, and a link to richer detail.

Well there, you really do learn something new every day (and i didn't mean the fact that RSS was XML, duh :)).

On the plus side, Atom looks like it's a far more well thought out for Syndication. I'm going to try and spend some time over the next few days reading up a lot more on it, and pondering what Atom could do with SIP. I'm sure there are some pretty funky applications out there (SUBSCRIBE to Atom feeds, and get NOTIFY'ed on changes anyone?)

As so many people are asking me, here's an update:

Jordan is being induced on Friday. We have 7:30am (!) appointment at the hospital. Hopefully things will be over fairly quickly (although it did take 3 attempts over 3 days with Demi), so we should have a Solaris Zourzouvilly sometime on Friday.

Jordan's pretty damned fat now (err, 'Large') and waddling around in the heat like a trooper, bless her. She's doing great; especially in a house with 4 flights of stairs!

Although i heard of it a while ago, i’ve just had a play with Windows PowerShell, previously known as Monad/MSH on XP, and I have to say – what an excellent concept. I’ve always been a big fan of .NET, and the thought of having the entire set of .NET libraries available to me at the shell is just – wow.

Take a peek at the PowerShell team’s blog for loads of examples of msh in action.

At this rate, i’m not going to have any excuses left for sticking to using a KDE desktop. With a decent shell finally on the horizon, it looks like the day i start using Windows on the desktop again may be closer than ever.

http://somafm.net/tagstrance.pls

See somafm.net

I noticed the following text on DCGLUG the other day:

DON’T BUY WINXP!!! Just cos its not supported doesn’t mean it’ll stop working!

I’d just make sure you back it up properly ( to your linux box) then re-install if it gets compromised

Now, this got my attention. Why on earth would someone give advice to continue using an operating system that is not supported. The security risks are insane. There are still millions (apparently, 78 million) of people using Windows 98 + ME – a ripe target for spammers, hackers, and organised internet crime groups who need zombie machines to perform their dispicable nufarious crimes. The advice, surely, should be to move away from Windows 98 and ME now that there are not supported as quickly as possible. My reply was as follows:

Sorry, but what amazingly bad advice.

Sure, you don’t have buy Windows XP, but that doesn’t mean you need to continue using a buggy, security hole ridden OS.

Using the age old car analogy: Would you advise someone to continue using a car without a seatbelt? assuming they were not a legal requirement, of course.

If you want to stick with using windows, then DO move to XP as soon as you can. Alternatively buy a mac, or move to linux. even if you move to win2k, fine. just whatever you move to:

-------> KEEP YOUR MACHINE UPDATED <-----

... and this is why you should no longer use 98 or ME. because there are no updates. People can now exploit it, safe in the knowledge MS will not release any more updates. Absolutely perfect for spammers, scammers, and hackers.

Even better for them, it doesn’t come with the “seatbelts” that 2k and XP now do.

If you do continue using win 98, then do not connect to the internet (for all of our sanity), or expect application upgrades to continue working for much longer. If you do continue using it, every time you get a piece of spam, then remember: people whole followed this advice are one of the reasons you are getting that piece of spam in the first place. If you suffer from credit/debit card fraud, then remember: the advice to continue using windows 98 was almost certainly the reason (or your lack of common sense).

Banks are 100% right to not support people running security risks to use their services – it encourages people to upgrade, something that is really badly needed.

However, it appears that some people will allow their hate for Microsoft get in the way of security, one reply looked like this:

No, no, no! Once again do not move to XP. If you have programs that need to be run on windows for whatever reason then you can run them reasonably safely off the net or behind a firewall.

If you are unfortunate to have to use IE for managing a bank account that you have no control over you have your old copy of 98 and configure your firewall to only allow that machine access to the sites/ports necessary to make the bank account work.

What you do not do is give large quantities of money to a company that makes the same mistakes over and over again – probably deliberately so mugs will upgrade.

Bizarre. This thread quickly turned into a handbag bashing argument between Linux Zealots and and me – and i’ll cover the points made in another post.

Someone with more sense pointed out that machines running 98 will probably not handle XP. A completly valid point – sort of.

However (and this is text adapted from this post i made), kettles, toasters, washing machines, microwaves, tumble drier, tumble driers, lawn mowers, and fridges to name but a few household appliances that all need replacing after some period of time/usage – generally, the price dictates the quality, and thus how long it’s usable life is.

While my grandparents had a washing machine that lasted 25 years, my mother gets through one every 2-3 years (5 kids!). I’ve only got through 2 kettles this year alone – but i blame the amount of coffee i drink. my mother has had just one kettle in 4 years… or at least one that looked the same last time i was there.

Explain why computers are any different? People seem to forget that windows 98 is now 8 years old, and ME is 7. A very long time in computer terms. 6 years before that we were lucky to have DOS, let alone windows.

i’d think you are a complete idiot for putting the public at risk if you were driving around a car lacking basic safety features. even more so if you were actually advising the use of dangerous cars. again, exactly the same thing with computers.

Cisco’s NAC stuff (incl 802.1x to an extent, too) does this already. It checks the software you’re running on your computer is up to a certian version (or patch level). If you’re running old software that is a known security risk, then your connectivity is either sevierly restricted (i.e, outbound HTTP only to allow upgrades), or not allowed at all. Obviously the not allowed at all thing is a bit silly in this case, as how are you supposed to upgrade :)

I’d personally like to see ISPs require something like this too. Why should Internet users who keep their machines up to date have to pay for the users who don’t?

Exploited machines sending out spam, acting as a DDoS drone, or trying to exploit other machines are using more and more bandwidth – which obviously pushes the costs up for the ISP, who in turn has to pass costs on to the user.

Not to mention that any decrease in exploited machines means decreases in spam.

Although it would obviously be possible to work around any sort of NAC that an ISP used, the idea would be that it’s the people who don’t know how to work around it that need this kind of access control in the first place. Generally, technical users are fairly good at keeping machines up to date.

I’m heavily up for computers requiring a certain “software safety standard” – for both the software and the user (the hardware already has in form of CE in Europe) before being allowed to connect to the internet. Hopefully a law will one day require this.

For gods sake, just make sure that any NAC methods implemented are open standards.

On a similar note, i’m becoming increasingly tempted to not allow users running known security risks such as Firefox pre 1.5.0.4, and unpatched IE versions to access our online services. The only thing i’ve not worked out yet is how i’d handle older browsers with backported patches as used in various linux distributions.

Argh! Jordan's now 4 days past her due date, and still no baby. She was 2 cm dialated on saturday, but not had a single contraction since. Rugh.

On a completley seperate note, i'm lookign for someone to take some load off my sholders due to my scarily hectic work + home life. Job spec here

Our SIP stack's dialog FSM needed an overhaul. It was time for proper PRACK support and do something about those really broken UA's with a totally messed up idea of what the SDP offer/answer model really is.

So i sat down and got through a lot of paper re-designing the INVITE FSM, and in doing so massivly tidied up the design of it.

Having a design both in my head and on paper, i started to proof of concept it.

Suddently things got even more difficult. A complicated FSM with sub states, events with different arguments, storage within for the states, and intra-state communication is damn difficult to implement in a non code-repetitive way (i.e, with a switch(x)/case) and clean implementation.

I spent a while looking at boost::statechart i had the basic functionallity. however … our SIP dialog library had gone up in size almost 400k!

Not only that, but the code was damn messy, and error prone with maintaning.

So i decided to try and solve the problem myself. And i think i have – at least for simple FSM’s

Here it is.

It’s only a first draft/proof of concept, but looks promising. Feel free to use/modify/criticise – let me know what you think.

It is not thread safe, it is not re-entrant, and will never be either of those things at least without wrapping within boost:threads or such.

However, it does not use a virtual method, so is impressivly fast. A random test of processing 1,000,000 events with 15 states jumping between each state took just over 100ms on a xeon 3.2ghz. Yum.

Todays rant is going to be about the ITU. A “standards” organisation that have been annoying me for far to long.

For the last year i’ve been working flat out on a fairly large C++ project based around SIP. One of the elements involved is a MEGACO/H.248.1 stack for use by our media gateway and controller.

The ITU published the protocol specification under a document named H.248.1, and the IETF has published it as RFC3525, naming the protocol MEGACO. The ITU provide a pretty PDF version, and the IETF produce a dull but error-free RFC for it based upon the H.248.1 document.

But that’s where the similarities end.

Because it’s mainly an ITU protocol, the ITU have developed specifications for additional “packages” that extend the protocol for features such as playing audio prompts into a context from a file located on a HTTP server, dynamic tone generation, and adaptive jitter buffers among other things. Without some of the additional packages defined in other h.248 series specifications, MEGACO is pretty useless.

Now on to what’s been annoying me.

From the ITU’s website: “ITU-T’s mission is to ensure an efficient and on-time production of high quality standards (Recommendations) covering all fields of telecommunications.”

Yet, to downlaod their apparent “standard” “specification”, i need to pay for it. about £1000 in total for the H.248 series documentation.

Maybe it’s just because i’m from the open source world, but it seems like absolute insanity that i need to pay around £1000 to be able to read a specification for what is supposed to be an industry “standard”.

Maybe it’s because i’m of the view “standards” should be available for all to read.
Either way, It makes me wonder what would happen if the IETF suddenly decided that all RFC’s are no longer freely available, and you need to pay 60 swiss francs each time you want to read one. Ohh, and lets charge to accessing drafts, too!

I’ve yet to find out what their excuse for charging for accessing standards documents is, and i’d love to hear anyone who knows why they think there is nothing wrong with what they are doing.

Ohh, and H.248.1 is a really badly written protocol specification, too.

And it’s a stupidly designed protocol.

And no one seems to have implemented it properly.

And the specification itself is buggy.

And there is only one reference stack around – written in erlang.

I hate MEGACO.