I am a software programmer working at RealVNC with many interests including singing, and am a Christian. Many of my friends remember me as a Part III mathematician at Peterhouse. Find out more or contact me!
Too many specs… some good progress on the OpenID though; all hammered out and ready to go with it.
Piles of ideas flying around my head:
* Best architecture and a draft protocol for a distributed cloud desktop system. Uses a DHT and spins up destops for users. Pages of scribbles…
* Just finished the spec for an RFB interpretation of OpenID. Only needs server cleverness! Key is writing a protocol for a server to forward key components to a broker. My version now works with a Relying Party behind a firewall or NAT, via a broker. Hope to get buy-in for hosting one at www.realvnc.com!
* More rounds of tweaks to RA3, my new security type for RFB I’m pushing. Addition of rekeying phase drafted to solve concerns over known-ciphertext replay after session-hopping code from #1 above implemented.
* Brian gave me a long explanation of some clever algorithms we use for our hosted services. Seem awesome and cutting edge.
* Back to considering how to gather and propagate peer adjacency information for tricks like WOL auto-magic. Ultra-distributed wide-area Bonjour wakeup! Like Apple but potentially better.
“Help, old SCSI disks for my 90s pSeries cost a fortune and things keep failing because we’re out of space. I keep pruning but it’s not enough. df shows I’m out of room everywhere.”
Never fear, you probably have loads of space. A default install allocates about half the hard disk. Run “lsvg” to find some free “PPs” you can dynamically add to volumes (like partitions, but logical rather than linear slices of the disk). Then “chfs -a size=+1G /opt” will sort you out.
I felt so silly after discovering that today. A bit like the day I found I could roll halfway through a TL upgrade because it mirrors your disks and keeps half the space aside for fancy tricks like that.
Waaaah!?!?! They had the best solution in the world for content tracking and deduplicating (so fair enough, they consumed vastly more resources in serving Reader than any RSS competitor…). Every time something was reblogged or reposted or copied, they’d manage to link it to the original post in a rather awesome way. I used to use their API to get a persistent ID for a post, knowing that their reference to the content was probably the most canonical of any on the web! Sad times.
HP-UX, HP-UX, we all love HP-UX. La-la-la-la-laaaaa! This issue is totally unbelievable, but why am I not surprised?
An interesting collection of comments and rants by famous people
A delightful induction. Seems familiar from *Intro to Algorithms*?
An interesting overview of various takes on MAC. I keep messing with SELinux or various sorts of process restructuring with contained capabilities for our product, but never manage to make significant gains. Pretty hard for a general-purpose commercial product with daemons to achieve much by cooperation! I feel type-based systems and traditional privilege separation with IPC are ultimately most helpful, if we were to try to usefully reduce our surface.
Links and helpful comments appreciated!
you lie in bed for hours in the evening or morning, inhabiting the intermediate state between sleep and work: reading BSD source code, but on your smartphone.
Just watched the first minute of each of these, as relaxation. They are hilarious! I couldn’t pick between the Muppets, Battlestar, and the slide whistles.