from a Yahoo! and XML geek

Quick Links: Consulting | Book info | Brain Attic | Home

Push Button Paradise

Micah Dubinko

Mon, 10 Oct 2005

Delicious

I finally started using del.icio.us. You'll probably be either amused or appalled by what took me so long: the name. I seem to be completely incapable of remembering where those arbitrarily-placed dots go in the name. Ending is with .us is a clever hack of the domain name system, and it's spawned a host of imitators, but that other dot really threw me.

Yes, naming is that big of a deal to me.

I also used Uche's PyBlosxom (ug, another name) plugin to make the things post regularly to my blog.

So my need-to-post-stuff-from-work problem is partially solved. The remaining challenge is that I consider the plain-text files on my personal machine to be the real or "canonical" version of, well everything. With this change, I have a subset of files that get created on-the-spot on my webhost, and never exist at all on my personal machine. However, I thing the same techniques I use to blast the files local->remote will work fine in the opposite direction for this one case. Or, since the data all exists on another site already, I could just forget about it and refetch the data if I ever need it. Who do you trust? -m

posted at: 00:44 | under: 2005-10 | 1 comment(s)



Hey Micah - I have found darcs to be the perfect solution for the exact purpose you are trying to fulfill.  If you want some extended samples ping me and I will point you to a few things that Kurt and I have worked on recently that should showcase quite well just how easy and, more importantly, how well suited darcs is for maintaining linked repositories.  In fact the more linked repositories the better as this then instantiates a file linking system which ensures that, no matter which file you need, it will be available from any location, darcs is constantly optimizing each repository such that as few actual bit consuming files exist, the rest are simply links to another node within the system.  Naturally this would lead you to believe that high potential exists for problemlematic access to files from any given node at any given time.  However, darcs learns the "habits" of the system and "plans" accordingly, linking to files without actually creating a local copy once the odds of problems have proven sufficiently low enough to justify the link.... at least, this is how I understand things based from the project docs and wiki... in reality I havent been running any of my darcs based repositories long enough to know for sure just how efficient this system is.  But with as many people as I know that have had enough time to analyze this and claim to have never had any problems it feels safe to assume the same will continue to be true for anybody who properly utilizes the system the way it was designed; which by the way is very natural and very easy to learn.

Hope this helps! see: http://darcs.net for more info.
Posted by M. David Peterson at Mon Oct 10 08:11:03 2005


Syndicate: RSS feed

What am I reading?
Don Quixote
Self-Editing for Fiction Writers
The Complete Joy of Homebrewing
Analog magazine
Compilers
TAOCP


What am I browsing?
BlogFour
Blake Ross
Brianstorms
Caveat Lector
Claus Wahlers
Copia
Cringely
David Temkin
Dave Hyatt
Groklaw
Mark Birbeck
M.David
Miguel de Icaza
Mitch Kapor
Norm Walsh
Omar Tazi
Sean McGrath
Sjoerd Visscher
Ted Leung
Tom Bradford
Wil Wheaton


Archives:
Link

Powered by PyBlosxom