Mostly technology-centric, this time.
IETF I18n Gasp · Somehow I missed this; John Klensin and Michael Padlipsky have an Internet-Draft entitled Unicode Format for Network Interchange; gosh, if the IETF acknowledges that the Internet needs to treat all the world’s languages as first-class citizens, the next step might be to deduce that the same applies to its documents, the things that describe and define said Internet.
Wrangling Solaris and Rails for Twitter · Joyent runs Solaris and hosts Twitter and Twitter is overloaded; so they brought in the Solaris heavy artillery and learned some weird things about Rails (patch already committed by DHH). I’m bugging the people who were at that meeting to write it up in deep technical detail, there are lots of good lessons in there for both the Rails and Solaris communities.
Reverse-engineering REST · Uh, Google’s showing off code to observe REST applications and induce WADL descriptions? That feels deeply weird to me and I’m too busy to deep-dive right now, but it might be important. [Update: A colleague at Google dropped me a note to point out that this is from a former intern with ambitious naming ideas; its name now doesn’t include “Google”.]
Brain Regain · This piece called India Grows Up tells the story of how the Indian entrepreneurs behind Riya decided to move it from Bangalore to the Bay Area. Anomaly or trend?
Correo · Hey, look at Correo; it’s trying to blend Camino and Thunderbird technologies and come out ahead of Mail.app, which feels like a very hittable target to me.
Comment feed for ongoing:
From: Market forces of Sour Grapes? (Apr 28 2007, at 23:40)
On Riya: Munjal Shah has been complaining about wage inflation for some time now. I run a small startup in Bangalore, and Riya hired one of our guys by more than doubling his salary... and our salaries were way above median.
Call it the forces of the open market if you will, but find me less than sympathetic to find Mr Shah hoisted by his own petard.
Longer term however, India needs to be a well thought out strategic decision, executed by people who understand the market and its trade off's. Ill conceived plans formed from a vague sense of cost advantages, or reading Friedman, are likely to end badly.
[link]
From: Evan Jones (Apr 29 2007, at 06:05)
Side note: REST Describe is "not an official Google product", according to the author. To avoid confusion, he has removed "Google" from the title:
http://blog.tomayac.de/
[link]
From: Damian Cugley (Apr 30 2007, at 03:48)
While it would be nice to be able to spell Bill de hÓra's name correctly in RFCs, I can understand the IETF wanting to be very, very cautious about creating a successor to NET-ASCII. It is only through strict adherence to ASCII that the Internet has been as interoperable as it has been, and HTML's use of ISO 8859-1 as a default encoding seemed like looking to the future at the time but now is an inconvenience because it means we have to keep referring back to an obsolete character encoding: now UTF-8 is here, ISO 8859 is beginning to look as out-of date as ISO 2022. I'm glad we don't need ISO-2022-savvy decoders to read RFCs.
Before permitting the use of UTF-8 to encode RFCs, the IETF would probably have to agree a subset (profile) of the Unicode features that can be used. For example, can RFCs use RFC 2482 language tags? They are required to correctly represent Japanese and Chinese text in a document (it affects how some kanji/hanzi are drawn) but none of the software I use is aware of them so far as I know. There might be a call for restricting RFCs to a subset of the repertoire that is known to be widely supported. THe haggling could take years.
The section on the draft about their discomfort about basing standards on Unicode when the Unicode standard itself has to be in a permanent state of flux makes for interesting reading.
[link]
From: Blaine Cook (May 01 2007, at 18:50)
The Twitter / DTrace write-up is coming. Coachella killed the possibility of getting any non-essential work done last week, unfortunately. ;-)
[link]