Journal TTK Ciar's Journal: House Hunting, MBT Wikification, Bye-Bye to MPI 7
Wow, my previous attempt to keep up on this journal failed miserably.. five months since my last entry! I just haven't had time
The most wonderful news in my life right now is that my wife (cobalt, aka invisiblecrazy) and I are finally shopping for a house! We are looking for something in Sonoma County (about fifty miles north of San Francisco) with a large lot. An acre would be okay, two acres would be great, and more would probably be a waste. I'm hoping we will find something in an unincorporated area (outside of any city's limits). Our current rental is in an unincorporated area, and we're loving it. There aren't many neighbors, no city laws (just state + county), no police (just a sheriff), a nearby well for water, and plenty of space for our chickens.
Sonoma County allows landowners to build permanent structures of up to 100 square feet without a permit, as long as they don't have a poured foundation and aren't used for permanent human habitation. I'm looking forward to knocking a few of those together for outdoor workshops, spare storage, and a mini datacenter. cobalt's looking forward to expanding her chicken horde, putting a garden in the ground (as opposed to the wine barrels she's currently using) and getting some goats.
Some less wonderful news is cobalt's worsening health problems. Her organs continue to ulcerate, despite the medication, and we've switched to a more local specialist. Her old one was great at first, but sort of dumped us a few months ago ("you're cured! no, really! goodbye!") and we don't need to drive to San Francisco just to be given the brush-off. She'll be going in for some exams this week, and her new specialist may proceed from there to some of the more serious drugs. It's a little scary, because the potential side effects are pretty bad ("spontaneous central nervous system demylenation" is about as bad as things get). But we have to do something. She's in daily agony.
A friend and I have decided to implement some software of possible commercial merit, and sell or license it to interested parties (or possibly run it ourselves, and charge end-users directly for the service). Forgive me for not saying too much about it. It doesn't seem like all that hard of a problem, and if I blab about it someone else might say "Hey! Good idea!" and jam out their own version before we can bring ours to market. But it needs to be said here because, hey, it's significant.
I've given up on MPI. It just blows. I've started work on DVM ("distributed virtual machine"), my own message-passing system. I'm shooting for something a lot like PVM but without its drawbacks. We'll see how it pans out.
Lately I've been plowing ahead on the wikification of my MBT Resources website. I have a lot of new material, about 13GB in 180,000 documents, which I intend to "wrap" in Wiki pages and make available in a heirarchical organization which a select dozen or so trusted editors can update/modify, while everyone else on the internet is free to discuss them (every wiki page has its own discussion page). The bulk of the work thusfar has been the writing of a few thousand regular expressions. Software will look at each document's pathname and contents, and decide how to categorize the document. Then it will wrap the document in an appropriate Wiki page. A couple of days ago I wrote a mockup of a few such pages, and discovered that OddMuse's markup wasn't quite up to the task. After much pondering, I decided that the best way to go would be to rewrite OddMuse's ApplyRules function, essentially replacing their markup with my own. I'm trying to make it simple and powerful, and familiar to those with previous experience with MediaWiki (like Wikipedia) and BBCode. This seemed like the right thing to do because aside from the markup, OddMuse is a really good fit for my needs -- it allows three levels of editing (administrative, editor, and discussion), provides anti-spamming features via a blacklist which I can import automatically from other OddMuse sites, and is already familiar to me and implemented in perl. I've already modified it some to divide pages and administrators into "domains", so that different administrators can have authority over different subsections. I've fiddled with OddMuse enough that I think I understand where its bumps and warts lie, and I think after I've rewritten ApplyRules I will be happy with it. If I switched to a different wiki system, I'd have to learn its bumps and warts, and might find myself in more trouble than I already am. Anyway, my replacement ApplyRules is coming along fine. Now-dusty memories of compiler engineering are coming back and providing me with all the right solutions. We shall see how it goes.
Oh, and one more thing: I don't think I'll be writing about politics anymore. It's just too disgusting. This country's going to hell in a handbasket, and people are too caught up in it to break free of their dysfunctional misconceptions and change course. My plan is to find a nice cave somewhere tucked out of the way and try to ride things out.
-- TTK
A cave (Score:1)
DVM? (Score:2)
Re: (Score:2)
Yep, it will be open source.
I am implementing the dvmd in C now, with plans to write a libdvm in C and a Parallel::Dvm native perl module. If I ever get off my ass and learn python there will likely be a python module (a wrap around the libdvm) too.
The general idea is: each server in the virtual machine runs a dvmd, but unlike PVM's pvmd it is limited to dealing with information about DVM tasks running on the local system. The individual DVM tasks each have their own UDP sockets open for accepting mes
insert subject here (Score:1)
"possible commercial merit"?!? Possible?! This is going to make you and I very wealthy.
I found a project we may want to run with. I've downloaded a tarball and will investigate further...
poured foundation and possible commercial merit (Score:2)
I will likely do what my friend Al Wolf did, and use gravel, sand, and heavy wooden beams for a foundation. The idea is to dig down a little where your beams will go until you hit a layer of dirt that's harder-packed, then pack down sand and gravel on top of it. The beams go on top of that, recessed enough that their tops come about to ground level, packed to either side with more sand and gravel. Then the floor structure goes on top of the beams. It requires more fussing than a poured foundation, and m
Re: (Score:1)
After doing some research, it looks like there will be a few more applications to this project than i had anticipated (possibly with a bit of modification from our original goal), and th