I have some time to write this diary entry: I'm trying to
find a bizzare (well, they only come in this flavour) memory
overwrite in Noah
and instead of doing it intelligently I took a brute-force
approach: I've set a watchpoint in gdb. It slows down the
emulator to a crawl but no wonder: every memory write has
just become a bit more expensive.
Since I'm in a dictionary business now I'm looking for some
free dictionary data. There is WordNet, which I have already
incorporated, there is Webster 1913 edition and nothing much
else. Or so I though. There was this guy, Samuel
Johnson,
who almost single-handedly wrote a dictionary in 18th
century. I think it was even before Noah Webster did. Some
guy who e-mail'ed me and looks like he's even more of a
dictionary geek than I am (he has 7 of them, including
antique edition of said Samuel's dictionary) said that this
is his favourite. Amazing thing is that this dictionary is
now in public domain (albeit it's probably even more
outdated that Webster's). Another amazing think is that no
one has e-text'ed this. There is a cd-rom version of first &
fourth edition combined (which I've of course ordered
immediately from Amazon (mind you, I'm all for a boycott, I
just couldn't help myself and it's not sth. you can just buy
in your nearest B&N)). I've also ordered paper 1836 edition
from www.abebooks.com (one of the 3 sites I know that let
you find out-of-print and antique books). It was strange,
because the price of it was incredibly low $25 while there
were editions from around that time that costed up to $2500.
I'm puzzled. I'll see in a few days what I've actually
bought.
Which leads me to another thing: believe it or not, but it's easier to find a dictionary from 18th century than out-of-print book by a contemporary, fairly succesfull SF author. You don't believe me? Try to find "True Names" by Vernon Vinge. I've found only 2 copies at www.bibliofind.com while there was at least 10 different editions of Johnson's dictionary. I'll hold off, though. According to Amazon a new edition of this book with foreword from (among others) RMS is slated to be published in April 2000. And don't tell me that IT IS April. I know, god damn it! I'm not holding my breath, either. Originally they were to publish it in September 1999, then in January 2000 and now in April 2000. I no longer care and they've just lost a costumer. When it comes out I'll just spend a day (or two) at B&N and read the book in one of their comfortable chairs.
I'm in a full swing of updating Noah. Last three days
are
marked by addiction to download count at PalmGear.
Ah, vanity. As one could expect the number
skyrocked (around
1300 d/l) during the first two days when it was still listed
as "new" software on the first page. Now it's down to 150
d/l per day. I wonder what will be the long-term average. So
far the biggest complaint was that the database is too big.
It is (5 MB is a lot in the little PalmPilot world) but I
find such arguments kind of silly. After all in case of a
dictionary size translates into quality: better definitions,
more entries take more space. But who am I to tell users
what they really want? My task is to code and make them
happy. So I did. I was lucky enough to have a quite
reasonable list of popular english words and I've used it to
produce a 3 abridged versions of the original database.
Maybe this will make those people give me more that 2 stars
in reviews (all 3 reviews I got so
far are almost the same: great idea, too big). I'll release
it tomorrow. This will be 4th release during 4 days.
The next step will be converting Webster. This is going to
make my head hurt. The markup is a stupid html instead of a
markup that actually means something. I think it will evolve
into some kind of AI. I'll have to use perl (which I hate)
to do the first pass over the data. So far I've written the
converter in LISP but its syntax for regular expressions is
too cumbersome for the job and the job will mostly consist
of regex'ing.
I went to to video store today. Found "Pi" which was nice
since I've read that this is a good movie but couldn't find
it. Then I had a moment of doubt. There is a category of
movies, you know, those independent ones, that always get
rave reviews but do badly in box office (there is, of
course, an counter-category, excellent box office and
completely flamed). I tend to watch such movies a lot
(that's after I'm done with the "box office" ones) and it
happens too often that those movies are real crap. At least
as far as I'm concerned. That made me rent Go too - it's
less risky choice. I figured out that if Pi is a piece
of
thing that likes to hit the fan I'll immediately console
myself. I stayed for a while instead of going straight to
the check-out line. And that, let me quote Jet-Li here, was
a mistake. How can you resist renting Scarface when
you've
been wanting to see it for a month now? (well, the clerk
said he didn't like it, but is there a chance in hell that a
movie written by Oliver Stone, directed by De Palma and
featuring Al Pacino and young Michelle Pfeiffer can be bad?)
And what about
Fallen Angels
that you missed when it was being played in
a small theatre two months ago?
The end result is that I now have 4 DVD's to see and only
one weekend. This is going to be an overdose even for me.
Ah, and I ordered my TRGpro today. They say it should come mid-May. A hell-of-a-wait but at least I will be able to test Noah on some real hardware and not the emulator (which is quite handy during development, I must admit).
You shouldn't really expect me to say in my first diary
entry that this is my first diary entry. It would be so
common,
so uninteresting, so boring... Hell, sounds like me.
If you happen to own PalmPilot don't walk, run to download
the freshest version of Noah , my english
dictionary. It's the biggest kid in the town (based on
WordNet 1.6 database, which means 122k entries) and
amazingly enough barely fits in 8MB. It has a few rough
egdes but that's why I call it version 0.50 beta2. More is
to come (especially when my TRGpro arrives, so far I've been
developing solely under the emulator).
During the development I've developped (every english
teacher out there must be grinding its teeth already) a very
unhealthy attitude towards those [self-censored] who wrote
the PalmOS. It's not fun spending a night trying to work
around stupid limitations of list implementation (like max.
number of items is 2^15) and then trying to make all
versions of the OS feel happy about it (after all who said
that you should expect the same behaviour when you're going
from 3.3 to 3.5)?
New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.
Keep up with the latest Advogato features by reading the Advogato status blog.
If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!