Just before the New Year we accepted an offer on our house in Norwich, which had been up for sale since September. It was a bit lower than we'd hoped for, but the buyer was an investor and so we hoped it would be quick and not involve too much hassle.

How wrong we were. We finally completed on Friday, with the money hitting our account today (thanks to the Ulster Bank being a bit shit about telling our solicitor how much was outstanding on the mortgage). I'd become convinced that it was going to fall through, that we wouldn't be able to find a new buyer and that we were going to end up having to rent out a house that it wasn't convenient for us to check up on (due to the radioactive body of water between us and it). So I am somewhat relieved to finally be rid of it.

I've been asked about the v3 DG834G in the past (mine is a v2) and always answered that I don't know anything about it. Earlier this week Ben asked me about it and let me have a login to his. He also took pictures, which are on the OpenWRT DG834G page.

Some comments, based on this:

• The ADAM2 bootloader is different. Looking at the output of strings on it I can't see the checksum message seen in the v2 version. I also can't see the FTP server strings I'd expect. However the default config appears to specify an IP of 169.254.87.1 so it might be worth nmaping that when the router is coming up and seeing if anything is open.
• The wireless card is no longer on a mini PCI card. I assume it's still connected electrically the same way, but I guess this is cheaper for Netgear to manufacture.
• The CPU is slightly different, a TNETD7200ZDW instead of a TNETD7301GDU. No idea if this makes any major difference.
• The board appears to have JTAG and serial port headers easily accessible. And possibly even JTAG for the wifi.

I think it should be easy enough to get OpenWRT up and running on this (possibly someone has even done so already); someone just needs to hook up a serial console and do some prodding.

My last laptop had a serial port, but only when it was inserted into its bulking docking station. My laptop of the past 2.5 years doesn't have one. I used to spend a lot of time in datacenters wanting to configure devices over a serial console, so as a result I've been using USB to serial converters for some time. And I've never really been that happy with them.

My first device was a Belkin, using the belkin_sa driver under Linux. It was free (always a plus) and was fine for basic 9600 8N1, but didn't work with my eTrex GPS or my smartcard reader. I just assumed they were picky about what they wanted to talk to - I didn't need to use them often with my laptop and my desktop has plenty of serial ports thanks to a 4 port Moxa card.

I then bought a cypress_m8 device, as Simon had one and claimed to be happy with it. While again it was mostly ok for datacentre usage it didn't work with the GPS/smartcard reader. Also I'd sometimes see dropped characters especially when driving it at 115,200.

I just assumed this was all a limitation of the devices; I'd have said it was the fact they were cheap and nasty but I know the retail price of the Belkin wasn't that cheap. I asked around on IRC and people said they'd had no issues with pl2303 based devices. So I bought one from DealExtreme for a bargain price, figuring it was worth a try. And it was. It works with my smartcard reader and my GPS. It doesn't drop characters anywhere near as readily, even at high speeds. And it was cheap as chips. I've seen others complain about the pl2303, but it's definitely the best of the 3 USB to serial chipsets I've tried.

Dear Lazyweb,

What is currently the best GPS chipset suitable for walking? Something I can pair up with my E70 to get tracks for importing to OpenStreetMap is the sort of thing I'm looking for. I can find various reviews that suggest the SirfStar III is best for the low speeds associated with walking, but they're all at least a year old. Has nothing better come along since? What about the SkyTraq Venus 5?

I mentioned on my switch to Movable Type post that there were a few things that were hopefully going to happen RSN and that I'd talk about them if they did.

Well, one of them did, and Steve helpfully dropped me in it earlier in the week - I was granted write access to the debian-keyring. It's worth pointing out that while Steve did some prodding around this the process started quite some time ago; back in November James Troup (the other current keyring maintainer, and at the time the only one) contacted me regarding an offer I'd made to help out in whatever way I could. As a result I got involved in the keyring RT queue and did some basic triage and trying to point people in the right direction, where such help didn't require any keyring privileges. I also started thinking about how keyring maintenance could be shared in a trackable fashion. I made some suggestions to James and he was largely in favour with a few suggestions and wishlists.

I'll get into discussing exactly how it goes at a later point in time, but for the moment I want to get a better feel for the process and procedures to fine tune things. To that end I've been working my way through the keyring RT queue, and have removed quite a few keys of retired developers, as well as doing a handful of replacements for developers who'd lost or had their key compromised. There's still a few more tickets in progress and I'm trying my best to work through them in a timely manner - if you have an outstanding ticket and haven't heard from me then please do feel to ping it.