configure.in
check for Boost which I lifted pretty
directly from the one in QuantLib's configure.in. Still no new code in
RQuantLib, unfortunately.
This release is the third one based on Knoppix 3.6 and brings a couple of updates:
More information is as always available at the Quantian pages.
Nice to know that the long-meant-to-be-retired box acting as the static webserver took it all in a breeze,
Otherwise, the bittorrent tracker page at U of Washington rocks: 40 completed transfer in the few days since the announcements, compared to 89 for last month's 0.6.9.1, or the whopping 288 downloads of 0.5.9.4 which sum up to almost half a terabyte. Gotta fill the fiber ...
This release is the second one based on Knoppix 3.6 and brings numerous updates:
More information in the Quantian changelog once I get a chance to update it, and general info is as always at the Quantian pages.
The rep recommended getting another broadband router -- replacing this cheap 3com CM29220 I aquired via Techbargains.com with a Motorola 5100 or 5120. Any opinions on that?
Anyway, sincere apologies to anybody who may have tried to look for Quantian or this blog.
Two days ago, Cheapbytes sales informed me that they have 0.6.9.1 available as well.
I should note that I haven't been able to boot my laptops from USB, but that may well be a limitation in these not-exactly-brandnew laptops rather than in the method per se.
More recently, I noticed a 16 mb SD card for a whopping $2.99, shipping included, on the always help TechBargains site. So I ordered one, and got it today.
And lo and behold, it now works. I first tried to transfer the files via Card Directories, a little Palm app to place files anywhere on MMC or SD cards. Didn't seem to work, though.
After upgrading to the newest version of Power48, I sort-of learned from the README that a) the HP roms needed to be converted -- for which one needs to employ a Windows binary, and having the work laptop home proved useful for that, and b) that these thusly generated files needed to be installed under Windows too -- pilot-xfer did indeed fail.
And lo and behold, having done that, Power48 now works and I have three new (emulated) calculators.
It shows the wild swing in the traded price of the 'Bush elected' contract over this afternoon and evening ... up to the point when all prices went to 1.0, and back up just moments ago. Was that a denial of service attack? Server overload? Operator error?
To those have tried to access these pages, or the Quantian ones, my heartfelt apologies. The best I can offer for such time would be ... the Google cache. Sorry for any inconcenience. One day I may look into real hosting, in the meantime this will have to do, in particular as the last few gripes are all from over a year ago.
The race switched back to a 5k and 10k. Lisa did the 5k, and was quite happy with herself. I tried to run the 10k rather aggressively which didn't go to well overall. My mile splits increased quite a bit after mile 2, but the total of 44:35 is still not too shabby with a 7:11 min/mile pace.
libR.so:
--- rpy-0.3.5.orig/setup.py
+++ rpy-0.3.5/setup.py
@@ -54,7 +54,7 @@
RHOME = get_R_HOME()
DEFINE.append(('R_HOME', '"%s"' %RHOME))
-r_libs = os.path.join(RHOME, 'bin')
+r_libs = os.path.join(RHOME, 'lib') # edd 11 Oct 2004: changed to 'lib' for 2.0.0
source_files = ["src/rpymodule.c", "src/R_eval.c",
"src/io.c"]
if sys.platform=='win32':
which may be useful to someone else trying to update RPy.
Update: Greg just told me by email that the fix is in CVS too, so a new RPy will have it.
This picture is from mile 11 and marked the first meeting with Lisa and our girls. I was obviously still in pretty good spirits
So after some further testing, packages of R 2.0.0 are now on the Debian
servers and should be in unstable tomorrow. Because of the new internal
interface to R packages, older packages will not load. Users can either call
update.packages() directly, or wait a few days until we (as in
Chris, Doug, Rafael, Steffen and myself) get the various
r-(cran|bioc|omegahat|other|noncran)-* packages rebuilt under 2.0.0. We will
try our best to have them all updated by Sunday.
The race went well, and I ended up running considerably faster than during the previous half-marathons with a pace just under 7:20 min/miles for a total of 1:36:01.
Bottom line: If you use Windows, you're asking for trouble.
As they say, truer words have never been spoken... Other than recommending OS X in no uncertain terms, he listed a host of required add-ons to make that one dominant OS cope with its own lack of security.
make, no further config needed.
I sent it upstream to Koen, maybe it'll show up in a future versiom.
diff -ru afio-2.5.orig/afio.c afio-2.5/afio.c --- afio-2.5.orig/afio.c 2003-12-20 16:16:13.000000000 -0600 +++ afio-2.5/afio.c 2004-09-13 17:12:50.548515800 -0500 @@ -184,7 +184,11 @@ #include#include #include +#ifdef __CYGWIN32__ +#include +#else #include +#endif #include #include #include Only in afio-2.5: afio.exe diff -ru afio-2.5.orig/afio.h afio-2.5/afio.h --- afio-2.5.orig/afio.h 2003-12-20 07:59:42.000000000 -0600 +++ afio-2.5/afio.h 2004-09-13 15:17:51.700744200 -0500 @@ -477,7 +477,7 @@ #ifndef MKDIR int rmdir (char *); #endif -#if !defined (linux) && !defined(__FreeBSD__) && !defined(sun) +#if !defined (linux) && !defined(__FreeBSD__) && !defined(sun) && !defined(__CYGWIN32__) VOIDFN (*signal ())(); #endif int fswrite (int, char*, uint); diff -ru afio-2.5.orig/compfile.c afio-2.5/compfile.c --- afio-2.5.orig/compfile.c 2003-06-24 16:32:20.000000000 -0500 +++ afio-2.5/compfile.c 2004-09-13 15:03:44.532576200 -0500 @@ -210,7 +210,7 @@ * version; */ -#if ( defined(sun) && defined(__svr4__) ) +#if ( defined(sun) && defined(__svr4__) ) || defined(__CYGWIN32__) #include #else #include
This is a maintenance and bug-fix release over 0.5.9.3, yet still contains a rather large number of changes:
Slightly more informations are in the Quantian changelog, and general info is as always at the Quantian pages.
To cut a long story short, udunits code is a little old-ish and its Makefiles would need an overhaul. I have made a local package you can find here for now, and would like for someone to adopt this and make it a proper Debian package. The Makefiles needs a rewrite: only a static library is built, prefix indirection doesn't work completely, the manual pages go into the old locations of /usr/man and possibly more little things.
So the package is there, I'll put it into the next Quantian release, but I can't maintain it beyond that. This would make a nice project for a sciences grad student who actually uses NetCDF, and hence udunits.
Anyway, the Rggobi interface package for calling Ggobi from GNU R had a fails to build from source bug which only came up now when the amd64 crowd tried to build it (as this dependent package is a semi-step child in contrib). Since the last package release, I had tried to upgrade to the newer Ggobi 1.0.0-beta, but failed to build either the currently package Rggobi or the newer 1.0.0 of Rggobi against it -- so the old version will have to be it for a while. A friend has a saying for the peticular nature of these build problems: BOINJ, and I guess we need to make that BOICA now. And now, I won't reveal the joke via the blog.
Since then I spent last week catching up on a mountain of email, fixing a few bugs and upgrading a few packages, and doing a few little things like getting the bike back from its annual inspection. No real programming work yet other than adding a few items to the Quantian TODO list; not sure when the next release will be. Beancounter needs an update too, hopefully soon.
So I just edited the web form at gandi for the DNS of the .com address, sent a signed email to the 'change' bot at db.debian.org for the .net address, edited the configuration for bind, edited /etc/hosts, edited apache's httpd.config, and updated the IP address itself for the mail forwarder. Now all I need to do is wait for DNS to catch up :-/
Somewhat of a silly game, really. Maybe I should splurge for a real provider like Speakeasy?
While nominally only a maintenance upgrade over 0.5.9.2, it still contains a rather large number of changes:
Slightly more informations are in the Quantian changelog, and general info is as always at the Quantian pages.
The run was interesting: I was going pretty steadily at 7:20 min/mile until about mile 10 when I really needed to, um, go to the bathroom. I didn't quite get back to the same regular pace, unfortunately. Finish time was 1:38:43, for a pace of about 7:32, which is once again a little bit faster than the previous two half marathons this year.
This release contains a small but annoying fix to setup_beancounter which was mistakenly pointing to a version of beancounter in /tmp. Following a hint from CPAN Testers, a dependency on Fiance::YahooQuote was added to Makefile.PL. Lastly, a contributed script by Joao Antunes Costa is now included in the source distribution.
The whole column is very refreshing for the dismissive tone of his SBC comments. Cringely is a much smarter geek that I'll ever be, and he overcame SBC in neat ways. I'll probably remain scarred from the long troubles we had getting service after we came to Illinois ...
Many thanks to Diethelm for the new Rmetrics release 191.10057, and to the CRAN masters for including it in the archive.I have updated the initial packages that had been prepared for and included in Quantian 0.5.9.2, and just completed uploading them to Debian's incoming/ directory. As brand-new packages, they will have to wait the customary ten or more days until the ftpmasters insert them into the archive. Once that has happened, they will be apt-get'able from you favourite mirror.
In the meantime, you can fetch sources and i386 Debian packages manually (sorry, no apt-get support here) from
http://dirk.eddelbuettel.com/code/rmetrics/
Happy 4th of July, Dirk
Very, very nice to have this exciting set of packages in Debian, and to be helping along in its transfer from the land of Windoze-only.
It was a hectic trip: arrived late yesterday and caught a little bit of a BoF session before working some more on the slides til the early morning (ahem). Got up, registered, listened to Eric Allman of Sendmail fame on Spam, then to Martin on Debian and MIA -- while I worked on the slides (ahem). Then left the conference and met some Bostonian friends for lunch, back to the conference, listened to some other 'Extreme Linux' presentations, gave my talk and rushed back to Chicago. I should have allowed for more time because what I managed to see was all good. Also odd to know that there were other Debianers and not having met any...
This is a maintenance upgrade over 0.5.9.1, but still contains a rather large number of changes:
Slightly more informations are in the Quantian changelog, and general info is as always at the Quantian pages.
This release contains a small correction to the documentation for the
database connection password, a new test routine to check whether database
connections can in fact be made, and a hook from
setup_beancounter to actually use that test.
The Cheapbytes.com page also lists Mfg. Price $19.99. Being the 'manufacturer' here, I don't think I ever said anything about prices. Let alone offered to sell any. Oh well.
So burning dvds is a good idea, as is running directly from a linux partition
as outlined in the lilo booting
HOWTO. That said, there are of course always a few machines which simply
cannot be turned into dual-boot or linux-only. For these, we have a very
nice new solution: just drop the iso image onto c:/ as, say,
quantian.iso, and then all one needs is a boot cdrom (such as
the current clusterKnoppix cdrom
from which Quantian is derived), and then use
knoppix bootfrom=/dev/hda1/quantian.isoat the prompt.
Clean, simple, easy. Doesn't cause heatburn or pimples, and worked on my trusted Thinkpad T23, a machine from work which knows only win2k. The full writeup is now up in the new windows booting HOWTO.
The race was, um, shall we say, different? I guess after living for more than
a decade somewhere along the great lakes, I don't get much exposure to
elevation. Real moutaineers will laugh themselves silly about this chart
Anyway, I was hoping to finally break 1h 40min, and trying to stick to
7:30 min/miles which worked so-so until I got really tired after mile 10 and
more or less gave up on the idea. Only to then find that the last turn gave
rise to short stretch straight downhill, and with the clock in sight, and the
chance to make it, I finished as if I was chased by the devil himself. The
announcer saw me coming, looked up the bib number and tried to announce 'Dirk
Eddelbrrrrrgggh'. Not sure if then swallowed his tongue
And I didn't mean to gripe (other than about the useless mile
markers). It is a beautiful course all through the subdivision, starting with
the starter castles, and eventually settling down along the Fox river to
finish in 'downtown' St. Charles. I think I'll be back next year.
Anyway, it has been corrected and a new 1.6.4 is out, which is otherwise identical to 1.6.3.
Work on 0.5.9.* should hopefully commence in the next few days, based on Wim's recent clusterKnoppix releases derived from the new Knoppix 3.4.
As always, more info at the Quantian pages.
Thanks again to everybody who participated in the finance sessions at the recent useR! 2004 conference. During the discussions, the idea of a mailing list for R and Finance came up. Thanks to Martin, such a list has now been created and can be accessed via the pageSubscriptions are trickling in at a steady rate, let's hope we get this list to be a helpful and lively forum.https://www.stat.math.ethz.ch/mailman/listinfo/r-sig-finance
from which subscription requests can be made using the usual confirmation system employed by the mailman software. Everybody interested in 'Finance' (we will try not to be too picky regarding definitions) and R us cordially invited to subscribe. Also feel free to forward this message to interested colleagues.
I even cycled in from home, which added another 2 x 12 miles or so for a total of around 54 miles. Got lucky and arrived home just before a massive thunderstorm hit
The race started outside Soldier Field, went down the Lakefront towards Hyde Park, and back to finish inside the stadium -- which was neat. Running conditions were pretty ideal---coldish and cloudy---which helped with yet another personal best of 71:59 or more than four minutes faster than the Lakefront 10 Miles race last month.
(And no, I cannot distribute the two packages described in the main talk as they were done at work where open source releases are not yet an accepted methodology. Hopefully one day.)
I should stress that this race is really well organised. Tons of bands along the route, lots of spectators, lots of water stations, and supposedly 30,000 runners registered. The lap on the Indy 500 tracks was somewhat silly -- no wind at all, and hot. Incidentally, that was my slowest mile split. Overall, it was bit excessive to drive 190 miles each way just to run 13.1 miles, but heck, with the weather that nice, it was a fine day off.
By the way, I'm really quite pleased with the Tungsten C. It feels less sluggish (400 MHz Strongarm cpu), it has ample memory (64mb), the screen is terrific (but harder to read outside) and the 802.11b wifi is neat. Drawbacks are the above-average weight, the fact that there are way too few open hotspots (but it's nice enough not have to go upstairs to check email) and that documents to go and the adobe pdf converter require windoze. So I only get to use that at work ...
I guess it has been almost a decade since I saw him last back in France, but he clearly is one of my all-time favourite musicians. The repertoire tonight spanned material from his awesome 60s recordings on Blue Note to the newer material from the 90s, and it was all delivered in such a lyrical way that is really unique to him. Seeing Wayne Shorter live was also neat; I don't think I've seen him since maybe 1987 or so in London. Gee, I'm starting to sound really old. Anyway, Hancock will play one more next week with Dave Holland and Jack DeJohnette. I should see if I can still get tickets...
Now better put some miles on before next month's half marathon
Which is too bad, because she is really good, with a very wide vocal range and unbelievable control. She did a few pieces from her previous albums as well as some that should come in the fall -- she recently signed with Sony and just finished recording a first album for them. Lisa got us awesome seats in orchestra pit -- second row, to the side. Close up and personal. Real nice, all told, and just too bad the audience didn't get more into it.
0.4.9.6 is a bug fix release with relatively few changes compared to 0.4.9.5, but has
As always, more info at the Quantian pages.
As always, Debian and CPAN have the new version; it is of course also on the local smtm page. Freshmeat will get updated in a short while. Feedback welcome, as always.
With Cebit coming next week, I'd suspect that the next Quantian release will be based on the Knoppix 3.4 that is to be released there -- but it may take a while until all clusterKnoppix and Quantian flow through.
As always, more info at the Quantian pages.
Debian and CPAN have the new version; it is of course also on the local smtm page. Freshmeat has been updated too. Feedback welcome, as always.
It turns out that this mightily confuses the AvantGo reader I am so used to. Despite several upgrade attempts from (gasp!) windoze and directly (nice hints are here and here (though you may need to be logged in) via pilot-xfer, it would sync nicely but never get the reader software going. At first I was confronted with 'AvGoPimPod1 not a shared library'. Deleting libmal and installing libsmal from the 'palm executables' bundle fixed that. But still no reader. It continued to babble about channels -- I guess that's their commercial product for intra-company news transmission.
Well, another attempt his morning fixed that. Deleting AGConnect, AvantGo and libmal before reinstalling them along with a new definition of the server in AGConnect did the trick. While the info is exactly the same as before, this seems to sail around whatever bug was blocking me before. I'm happy.
Now if only the other channels would use a smaller font just like the venerable Globe & Mail does so that one gets more mileage out of the neato 320x320 display...
Now based on kernel 2.4.24 (while the openMosix patch is being finalised for 2.4.25), this is still Knoppix 3.3 based. A new version based on Knoppix 3.4 will appear some time after CeBIT. As always, more info at the Quantian pages.
Back now at the computer, I just saw MJ Ray's riposte which I happen to, no surprise here, agree fully with. To me, one of the nicer aspects of Debian Planet is how it, at least occassionally, brings out the other, non-computing side of fellow Debianers. That is clearly not a bug but a feature.
Coincidentally, while running errands this afternoon, I listened to Chicago Public Radio (one of the very few remaining stations worth turning on) which carried an interview with a U of Chicago law prof on the origin of the word 'echo room'. His thesis is that people self-select themselves more and more into virtual communities. The main point here that these communities are all non-overlapping, so people simply reinforce prior beliefs and opinions as the only feedback they get is from people who, by virtue of the pre-selected, are likely to hold views very similar to their own.
Getting back to the prior topic, having non-Debian posts on planet.debian.net allows us to break the 'echo' pattern at least a little bit.
This fixes only a minor thinko in which GetConfig() was exited (apparently)
too soon if no ~/.beancounterrc was found, and should be of
interest most to very new installations.
Right now, and for kicks, I manually turned html mode on and off again as we don't formal document header for a blosxom input file. Better than nothing, but far from ideal.
And silly at the same time -- as it is about quantifying the arguably rather multidimensional aspects of, gasp, several social sciences discplines onto a two-dimensional plane (spanned by the, as far as the site goes, outdated 'left / right' metric, applied to an Economics axis, as well as polital liberalism vs authoritarian axis). Given that I was nerdy enough to endure schooling all the way to a doctoral degree in quantitative methods in social sciences (Econometrics, if you cared to ask), I have quite some empathy for the undertaking (as I also care a lot about empirical work if and when it is done well), as well as countless objections. But then again, we just did that mostly for the fun of it, didn't we?
And no, I'm not going to say how I scored on these two dimensions of
'economic left/right' and 'authoritarian / libertarian'. But if I told you
that I've read the Economist for close
to two decades, you'd get an idea.
The main new feature is support for the lightweight and fast SQLite database backend (while PostgreSQL, MySQL and ODBC are of course still supported as always). A few minor bugs were fixed, and several additional documentation files were included for good measure.
Now available on the beancounter pages, and should soon be on the Debian and CPAN mirrors.
Both these oversights have now been corrected.
So I built those for him in December, even stuck them into the 0.4.9.2 release of Quantian I had prepared about the same time ... and then never uploaded them. Ouch. Shame on me, I suppose.
Steffen, being a nice guy, sent me a kind note today that he had prepared a new package based on a new upstream release of qtl, and gently reminded me of the RMySQL and DBI packages. Blush. Well, they are now on master, and even rolled RMySQL forward by a minor release. Or is that considered uncool? Another brown paper bag?
As a new maintainer for blosxom is looming (Salut, Pascal :), I tried to
clean the package up and switch to dpatch ... only to fail
miserably. I had converted another package of mine to the 'simple patch
system' of cdbs, and that is soooo much easier to use. Oh well.
Shipment was supposedly to be in three days from now ... yet the box arrived today! It boots Knoppix as well as my Quantian just fine. A pIV 2.8 GHz and a mobo with sata support, gigabit lan, graphics, sounds, whathaveyou all integrated along with a puny little disk and a laughable amount of Ram (order for a gb from crucial is on its way), it all came to just over $500 of which $100 should come back via a mail-in rebate.
So I quickly added three more lines to call tzset and tzname, and now assign
a tz variable which can be used in the rss format. One minor annoyance is
that the setting itself is still in the flavour file
/usr/share/blosxom/timezone -- so I dpkg-divert'ed mine.
If you see this via rss aggregation at planetdebian, it should show with a time of very early in the (UTC) morning whereas I type this at 21:00h Chicago time.
By the way, if someone wants to take over maintaining blosxom for Debian, drop me line.
Enjoy -- Once the mirrors are updated, you can get it from both Debian as a package or source tarball, the CPAN Perl archive or via its Freshmeat page. Feedback welcome, as always.
