Here's my CPAN goods. [cpan.org]
#pragmas and compiled-in stuff, but I suppose serious computational hackers need that.
It did give me notions about trying to hack Perl's iterator opcodes to provide auto-threading versions...guess I'll need to paw thru the compile sources, and try the midnight opcode swap hacks I've seen posted here and elsewhere
They gave us a flash drive w/ *all* the tutorial presentations on it, so I'll dive into the MPI stuff ASAP; maybe that'll be a bit more Perl-friendly (there are a couple MPI modules on CPAN, one recently updated).
Alas, my schedule this week won't permit me to attend the regular sessions, but they're mostly paradoxically brief: 30 minutes to discuss anything related to HPC seems a bit foolish, except for cocktail parties and Nova snippets.
However, I'm definitely going to the exhibit hall Wed. or Thursday. I was astounded at the number of exhibits/vendors, and the elaborate displays I caught a glimpse of from the convention hall lobby. I can only assume that there's serious $$$ in HPC these days - I'm guessing most of it is gummint money. Still, it does warm the heart of an old parallel coder to see new life in the discipline. However, based on overheard conversations, and the content of tech sessions and tutorials, I'm guessing the HPC hackers are still grinding out code the same way we did 15-20 years ago, which is a bit disappointing.
I'm still convinced there's a big opportunity for dynamic languages in this space, tho likely not for the serious computational stuff. And it resurrected a memory of something I'd been meaning to track down for a Thread::Sociable storage manager: RDMA.
As someone who's been working inside, on, around, over, and under MPP systems for 20+ years - I just picked up the book on UPC a few weeks ago for some hints on supporting DSM in Thread::Sociable - , this is really sweet, and likely to eat my week.
Wish I'd known about this before the early registration expired (arrrgh!)
Who knows, maybe it'll make the Strawberry Perl dist...
Update: New 1.02 release: Prior version did some URL munging it shouldn't have.
Yeah, me too. But I'm a capitalist, so I like to see what people are writing checks for. As I'm fed up with the various unsubstantiated claims, I've whipped up a little graphic that will hopefully cheer you up. I'll try to keep it updated regularly.
If nothing else, it'll give us all something to watch as we're overtaken by our Ruby and Python overlords.
Despite searching all over the 'net for a tool like this, the closest I could find was Aptana, which is Eclipse based and thus painfully slow and bloated (and kinda contradicts the whole idea of browser based GUIs, doesn't it ?). Why didn't I find this before ? Why didn't this pop to the top of my many Google searches ? Arrgh ! I've been splashing about with YUI, Ext, jQuery, Dojo, et al. for months now, crafting stuff by hand, and here was this wonderful solution that I've never heard of just sitting there.
Guess I need to get out of the house more often. But that won't happen for awhile, cuz I'll be wired into TIBCOGI like Neo into the Matrix.
PS. Also liked InfoWorld's choice of NetBeans over Eclipse. I used to think it was just me, but they've confirmed my suspicions: Eclipse's huge collection of plugins is killing it. I've had people recommend Eclipse as a solution for simple data analysis/BI tools...and then I'd stop and think about the BI users/analysts I've known over the last 15+ years. Very smart people, but if you dropped Eclipse in front of them, you'd lose a lot of friends and associates very quickly. And help make MSFT a bigger pile of cash on Excel licenses.
The recent arrest of Senator Craig has undoubtedly provided a significant boost to your efforts to host future political conventions. I have no doubt you're updating your convention prospectus material as I write this. The rallying cry of "Never fear: no queers here!" will undoubtedly warm the hearts of convention delegates of both parties.
However, if I may, I'd like to offer a few suggestions to avoid potential future unsavory episodes:
I hope my suggestions do not fall on deaf ears, In this moment of glory, full of a sense of pride and self-congratulation, you and your constabulatory may look askance at the "low hanging fruit" available to control the more salacious nenbers of your community.
Until later,
Tab, Your pacifist Working Boy
There was a quick speech that we're all biased suckers. So stop doing that.
Then the best: the leader of the Swedish Pirate party, gave a very coherent and direct message about why copyrights suck, and why we should be mad as hell and not take it anymore. Closing speech by <I forget> about branding open source was very entertaining despite technical glitches.
wxPerl: Alas, more about Eric's app than about wxPerl. I didn't have to give them grief about the lack of docs: someone else did.
Error Handling in AJAX: O'Reilly goofed! Speaker was a no show for an SRO talk; turns out he'd told them months ago he couldn't make it. Grrrrr...so I sat thru a few State Of lightning talks. Nice to see PostgreSQL making such good progress.
DBI BOF @OBF: Went very well. Tim B, jZed, and myself only, but the weather was good, the beer was cold, and its not nearly as crowded on Thursday as Saturdays. jZed and Tim headed back after a couple hours, but I stuck arround and listened to some great jams under the Portland sun, burned a tasty 'gar, and did much sampling before heading back for the Perl lightning talks.
Lightning Talks; Pretty good, tho I didn't see the sort of spirited talks I've seen in the past (e.g., "Stop Using XML For Everything, Dammit!"). Before the Onion talk, a TPF auction was held, at which I picked up PBP, which I've thus far managed to avoid. Once I finish the Erlang book, I'll do my best to give PBP a good read (tho I'm admittedly skeptical of such tomes).
State of the Onion: More techie than spiritual this year (perhaps the focused audience reined Larry in ?)
Evening Festivities: Hit the Sourceforge party w/ jZed. Was OK, free cocktails, but not much grub. Then the night began in earnest when jZed took me down to Outlaws to see the March Fourth Marching Band. Yeehah, what a hoot. Danced my ample buns off...as did everyone else. Massive props to jZed for the invite. He is now my personal Lord and savior! GO TEAM!
Due to the late night, I couldn't drag myself back to OSCON for the Friday keynotes. So now I load up the Jeep and beat feet to Seattle for some R&R, including a ballgame at Safeco, before heading back home.
This was probably my favorite OSCON of the 5 I've attended. Very meaty, with extra crunchiness of the parallel processing variety. Too many ideas to pursue, too many books to read. But entertaining and educational. Kudos to the O'Reilly crew.
Sat thru Greenplum's talk. As someone who's been playing in the MPP sandbox for 20+ years, the stuff they're doing really intrigues me. (Think API for automatic parallel decomposition over petabyte datasets). I may have to hit Luke up for the job I turned down a few years ago.
Then Xquery+XForms. Alas, the speaker seemed unaware he only had 40 minutes, so much of the meat got truncated. But I need to revisit the whole XQuery thing I've thus far dismissed.
Nested data parallelism was interesting, tho being Haskell specific, was perhaps a bit impractical. However, it does dovetail nicely with the Greenplum talk; I wonder if they've looked into it ?
Then a quick talk on YUI CSS by webchic. Definitely gotta add YUI to my toolkit.
I then collared Simon Peyton-Jones for a quick chat to help get my mind right about the stuff I'm doing for STM in Thread::Sociable. Lots more to read and (I fear) to implement. But I think I've got the basics right. I can cheat on "orElse" operator with a simple boolean "acquire()" method; not perfect, but usable. I/O remains a challenge, tho I've vague notions about writing a PerlIO layer to support thread ownership and buffering for such things. That will likely wait until the 2nd or 3rd release. A transactional queue may bridge the gap for tte time being.
The Javascript Performance talk was a high-speed adrenaline fueled affair, with conclusions I had come to already (i.e., Dojo/Scriptaculous/et al are great, but too fat...so get out the trimming knife), and also some new ideas (remove the DOM head before hacking up the DOM...gonna have to try that with those monster PPI::HTML[::CodeFolder] documents).
Lastly, the Machine Learning with Perl talk was far too brief. Great insights, but not enough meat. Would've been nice to see more on how he used PDL.
Looks like Thrusday's DBI BOF @ OBF is a go. Should make the later lighting talk sessions more enjoyable.
Data Mining Open APIs: much more about data mining than Open APIs. Note to self: never eat the turkey sandwich for lunch before attending a 2 hour talk on statistical analysis in a darkened room. The Python examples illustrated the computational brute force approach to deta mining. And seem completely devoid of code that manages the collection/cleansing of the data. Definitely *not* a solution for ad-hoc OLAP. But wrap the API access as a storage engine, add in some SQL::Statement, and all that code turns into a single SQL statement. Then we start JOINing between SOA's. And maybe toss in a bit of charting syntax for chuckles. Ah, if only I had the time. The best take away was the link to programmableweb.com/apis. Maybe need to scrape these sites and derive schemas (WSDL should be a nice start)
But the best part of the day was the excellent Erlang BoF. Very similar to Haskell (presumably due to the same functional programming roots), but battle tested for 10-15 years in systems with 6-nines uptime SLAs. While it was only an hour long, Patrick Logan provided lots of intriguing info (esp the auto-update of modules in a running system, something anyone working under rigorous SLAs can appreciate). I'm still a bit foggy about using immutable variables, but by happy coincidence the new Erlang book was available at Powell's onsite store. So I've got fresh reading matter for the remainder of the summer. Too bad there wasn't a full Erlang tutorial: the Haskell presenter was interesting/entertaining, but Patrick's "learn by doing" approach worked much better for me.
Day 2 promises intrigue (the semi-stealthy Intel announcement at the morning tutorial) and enlightenment: Data Mining Open APIs appears to mesh nicely w/ my efforts RE: DBD::Amazon and some other similar drivers I've been working on (albeit at a slow simmer).