So, here's what I'm trying to do. I'd like to read my email from two machines. I will leave all of my email on gmail's server and I'll just track the emails that I've downloaded locally on each machine, so I guarantee that I have a copy on each client machine.

This shouldn't be a problem using normal POP3 protocol, but gmail thinks that it's smarter than everyone else and 'fixes' the broken (inefficient) protocol by just remembering which emails are downloaded - even if they're not marked as 'read' on gmail's web interface.

I download an email on one machine. The email comes over, and doesn't get deleted or marked as 'read' on gmail's web interface. I try to fetch it from my second machine and it says that there are no new emails. Even if I specify "download everything, I don't care what it's status is" gmail says, "no, you already grabbed a copy of all of your emails, I'm not sending you any more copies of it". You can reset Big-G's counter by resetting it for all of your emails in the POP3 settings, but this resets it across the board. I could download the whole shebang to my second machine, but any future messages will have the same "one copy only" problem.

I know why Google does this. They do it to save on bandwidth and to make pop3 faster. The way POP3 works is the client will ask for a list of ids on the server if the client wants all messages - Google doesn't give all ids, it just gives un-fetched ids. This is not standards-compliant and should be fixed. They want to remove the possibility of someone using the "never delete anything" option, so when the client says "give me a list of all of the ids on the server", it doesn't do a full table-scan of the user's spool every 30 seconds that a client checks in.

If you have an account on badcheese.com and want to connect to my git repository, here are the instructions:

On your local machine:

mkdir ~/git-badcheese
cd ~/git-badcheese
git config --global user.name "Your Name Comes Here"
git config --global user.email you@yourdomain.example.com
git config --global color.diff auto
git config --global color.status auto
git config --global color.branch auto
git clone ssh://badcheese.com/git .

Recently, I bought an HP dv7t-4000 laptop. I've had it for about a week now and I'd like to give my 7-day review of the product after using it for a while. I'll start with the good and work down to the bad.

Wireless: (10/10) I can see more wireless accesspoints than I could even see in my home with a macbook pro.

CPU Power: (10/10) I upgraded to the i7 processor, so I've got a 4-core CPU with hyperthreadding (so 8 cores to the OS) in Windows and in Linux.

Memory: (10/10) Comes with 6GB ram. Has not been too little for me yet even with virtual machines running on it.

Screen: (10/10) Awesome 1600x900 screen. Very bright and clean-looking.

3D Graphics performance: (10/10) I opted for the ATI graphics card which performs nicely. I loaded up a couple of games and maxxed-out the graphics settings on them and I got a minimum of 15-20 fps, so I'm very satisfied with the graphics even though I don't do a lot of gaming overall.

2D Graphics performance: (9/10) Pretty nice video playback. I can do full HD video on youtube and I only get an occasional choppiness. I'm guessing that the CPU is not even breathing hard doing the decompression. The graphics card is pretty capable, so the choppiness could be the network or flash or something else. Not really all that worried about video - it works and it's nice enough.

Price: (8/10) The laptop retails for $1500, but I got it with an online coupon for around $1000.

Heat: (8/10) It runs a little hot. Not unusual for any laptop, but you could char your left leg a little if you put it on your lap cock-eyed or something for more than a few minutes. You could warm a soda in a very short while if you put down a soda near the vent. The vent is not hot-to-the-touch, but it definitely puts out some heat.

Here’s some first impression about my experiences with the Drobo.

I purchased the Drobo (v 2.0) with a DroboShare controller (embedded Linux box that exports the Drobo to your network).  It has a third-party NFS server on it but it doesn’t work very well.  I tried the Samba exporting on windows and it works great, but mounting via CIFS on Linux kinda sucks and has frozen a normally stable linux machine more than once.  However, the Drobo *does* make a fantastic rsync server.  It’s fast, it’s stable and you can do most everything over rsync that you could want to do with any other protocol except for edit a file or something like that.  It’s great for backups and great overall.  The theoretical full size of it can be 16TB, but I’ve got a 1TB and a 500GB drive in it now and it’s allowing me 500GB of usable space.  The logical volumes are 2TB, but I’m guessing that the Drobo will go into non-stable mode (and will ask me to insert another drive) if I exceed the 500GB that it’s got mirrored on the two drives.  I love that all drives don’t have to be the same size (take that ZFS) and that it tells me when a drive is failing and when it’s time to go out and buy a new/larger drive to add capacity.  So-far it’s very stable and fast enough to do most operations.