layout hack
layout hackMain Pagelayout hack
layout hack
RSS Feed RSS Feed Main page
layout hack
layout hack
layout hackAbout Melayout hack
layout hack
Software I wrote
Resume
Friends of mine
Pictures
Musicianship
Stuff I have for sale
layout hack
layout hack
layout hackPersonal Newslayout hack
layout hack
2010:
March, April.
2009:
January, March, August.
2008:
Jan, Feb, Apr, May, July, August, September, October.
2007:
Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec.
2006:
Jan, Feb, Mar, Apr, Jun, Jul, Aug, Sep, Oct, Nov, Dec.
2005:
Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec.
2004:
Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec.
2003:
Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec.
2002:
Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec.
2001:
Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec.
2000:
Jan, Feb, Apr, May, Jun, Jul, Aug, Oct, Nov, Dec.
1999:
Jan, Feb, Jun, Oct, Dec.
1998:
Jul, Aug, Sep, Nov.
layout hack
layout hack
layout hackGeek Stuff (computer related)layout hack
layout hack
Digital Music
Java
Why LiveWire Sucks
Why ASP Sucks (a bit)
Linux
MacOS
Unix
Oracle
Perl
Emacs
O'Reilly
layout hack
layout hack
layout hack(some of) My Interestslayout hack
layout hack
Humor
Sony Playstation
Cars
layout hack
layout hack
layout hackSearchlayout hack
layout hack

layout hack
layout hackAdslayout hack
layout hack



Valid HTML 4.01!

spacer
August 21, 2006

The guys on the DL.TV podcast are excited about the new Mac Pro because they have to compress the video of their podcast into many forms for download, which takes hours and hours, and the Mac Pro may cut that time in half, or better.

Generalize that example and you have the answer to the folks who say that nobody needs a dual-core CPU or more than 500MHz or so since we're all just typing and the computer is waiting on us. Well, we're not all just typing. Never mind the music we're listening to or the 40 windows with tabbed web pages in them that may have JavaScript or Flash code running in them; think about this: processing, storage, and bandwidth are not independent constraints.

Caching and indexing can use cheap storage to overcome bandwidth problems; examples include desktop search indexing (trade significant processing time and storage space to provide fast searches later, overcoming the storage bandwidth constraint), browser caching, IMAP message caching, and so on. This is familiar. But processing power can do something similar, as we've seen with the industry-transforming impact of something as nerdy as an improvement in audio compression. CD's are (or were, 10 years ago) too large to trade or store in WAV format, but a 10:1 compression ratio made it reasonable. That enabled Napster, which changed everything. Video is still unwieldy, and BitTorrent is both too complicated and too inefficient to enable a Napster-like phenomenon for video. But video codecs are getting better at maintaining quality while reducing file size, at the cost of tremendous processing effort. Zip and GZip and RAR are more obvious examples, but they don't enable things that weren't possible before. But what about Skype and Vonage? What if the audio isn't an MP3 but is an international conference call? The music distribution industry is feeling the shock waves, but the telephone industry isn't feeling this yet. It will. If we're just typing, except really we're on the phone and we're typing because we're taking notes, maybe the phone is really Skype, using our "excess" processing power.

The next industry to feel this pain will be the video distribution industry, by which I mean the movie and TV folks, Blockbuster, Netflix, etc. MPEG-2 is still too big and unwieldy to move around, and everybody is still stuck trying to figure out how to replicate the Blockbuster Video model using MPEG-4. One problem with that is that people barely know how to use their VCRs, and DRM encumbered downloadable video is much more painful to deal with than the videotapes and DVDs that Blockbuster offers. The other problem is that the video files are still simply too big. There either needs to be a widely offered, financially subsidized set top box that can play back H.264 or some other high quality video format (satisfying the needs for ease of use and download time targets), or the market will solve this problem in an open and peer to peer fashion, a la MP3 and Napster. Even if video codecs don't get any better, fast forward a few years and low end laptops and desktops will be able be able to encode standard definition H.264 in real time. Slap a nice user interface on that and every teenager is their own TV station. Better yet, every teenager is a news reporter. Every conference call becomes a videoconference, and call quality finally stops being so horrible because everybody's on a headset phone instead of a speakerphone, and their headset and VOIP software are doing echo cancelling and ambient noise reduction. Face to face interviews with a news van and camera crew aren't necessary if the interviewee has a decent webcam and phone. And so you're sitting there "just typing", writing down interview notes, and your CPU is working its butt off.

Fix the awful "thundering herd of incoming connections" problem with BitTorrent and it's Napster for video, and game over for the movie folks. And the thing is, the movie folks aren't pursuing convenience, nor are they interested in the sort of major investment that would be required to put their closed-system video on demand set top box in the same price range as a DVD player today. Right now, the price of a system that can do this sort of thing, using mass market hardware and free software, is still probably around $800 (based on this system built last year). DVD players cost $50. But storage and processing costs are decreasing very quickly, so it's just a matter of time before a suitable set top box costs less than an iPod does now. Maybe 2008? And that's without a hardware subsidy.

It'll probably fall to Apple, again, to sell something that's closed and more expensive than necessary, to provide an endpoint. (And by the way, I really don't see that endpoint being an iPod-sized device, or even portable. I think that whatever endpoint succeeds will be a set top box of some kind.) If they do that then they once again have the opportunity to control the DRM format and become the 800lb gorilla. But by the time that makes financial sense for Apple, it'll also be easy for the masses to swap unprotected video in a P2P fashion. Perhaps the only thing stopping that from happening now is that there isn't a single, overwhelmingly dominant video format to match MP3.

So, bring on the "excess" CPU power. It'll continue to make bandwidth seem cheaper, make previously bandwidth-constrained things possible, and make previously price-constrained applications (like having a computer capable of encoding two video streams in real time at the same time in your living room) possible.


August 17, 2006

My first CD is available for purchase now! Woo hoo!

I also have a record label web site for my new record label. (Yeah, I already know it looks bad in IE. I'll fix that later.) There are full-length low quality samples, and one full-length high quality sample, on the page for the new CD. CDBaby has high quality samples but they're only the first 2 minutes of each song, and most of the songs take a while to build up and really get cooking about 2/3 of the way through (and they're anywhere from 5 to 11 minutes long).


August 15, 2006

Last week I finished mastering my first CD and it's been sent off to CD Baby. It's not in the online store yet but it will be shortly, for $8. I started my own record company to publish it. I did the design and packaging myself. More details on all that stuff soon.

Last week, I realized that the preceding two weeks of occasional soreness in my hands (especially my right hand, especially after using the mouse a lot, especially in the parts of my hand that I use to move the mouse and the scroll wheel) was the onset of repetitive strain injury. Being a musician and computer programmer, that's bad. I immediately went to the Apple Store and CompUSA, and they basically have no ergonomic pointing devices for sale. A trackball that's exactly the same shape and size as a mouse and is used exactly the same way except for with the fingers instead of the arm doesn't count, guys. It's actually worse. There are dozens of nearly identical mice on the shelf, and nothing remotely close to the Evoluent VerticalMouse 2 and 3M ergonomic mouse that I ended up buying. I've used the 3M mouse longer since it showed up on Friday and the Evoluent mouse only showed up yesterday. Both are fine and made the pain go away. Both required some getting used to, but I like them better than the "just mouse with the left hand" solution since that means you also have to learn to use your right hand for modifier keys (Ctrl-click uses the other thumb now, etc.). I didn't get the SmartNav, but it looks pretty good for serious RSI sufferers. It seems like overkill (and a huge re-learning chore) to me.

Funny: Spamusement mocks Sex and the City.


August 14, 2006

Our photo gallery was offline for a month or so following my recent home server overhaul project, but now it's fixed. Just in case you cared.

I'm cleaning out my old news links, so don't be surprised if some of these stories are a few weeks old:

Via Tero: Opinion: Internet isolationism is bad for business. Okay, I sort of agree with net neutrality, though with all of the carcasses of businesses which have failed due to lack of micropayments, I wonder if net un-neutrality might be another way of saying that some services show up on your ISP bill instead of incurring the overhead of a credit card charge. But probably not, because telcos are too colossally stupid to think of something like that.

What is interesting to me is that everybody is talking about Google Video as the poster child of who would be hurt by tiered internet service charges. I can think of another service that's a more immediate threat to telcos. What does a telco charge you a variable rate for on your land line? And what's the biggest thing that businesses and individuals have been embracing regarding telcos and the internet lately? Hint hint.

Schneier on why data mining for terrorists can't work.

What if you made a captcha based on physical attractiveness? Have fun trying to judge hotness of "the other team"; I was clueless at picking hot males. Try it with your significant other and wait for the laughing at each other's cluelessness to ensue. "Eww, no way." "Oh yeah, totally!" Lotsa yuks.

Nerd note: when using Ubuntu 6.06 LTS, if you're using Terminal.app on Mac OS X and it says "unknown terminal type dtterm", you need to apt-get install ncurses-term to get that terminal type installed.

Via Faisal, a story on the homeless problem and how much we actually pay to keep these people alive: Million-Dollar Murray. I can't remember who the jackass was but there was some ultraconservative TV pundit guy who, about one day after I read this, went on the air to interview someone from Seattle who was basically trying to take the basket case types off the street and get them out of the cycle of gutter to emergency room. He claimed he was a recovered alcoholic and that the only way to cure people of alcoholism is for them to hit rock bottom. Strangely he seems to think that being homeless for a few decades, with the associated health problems, doesn't qualify. The article I linked to mentions that most people are homeless for a day and they get so scared that they do whatever they have to do to never be homeless again. Makes sense. But what about someone who's been living in a tent under the freeway for years, with no sense of personal safety or knowledge of where the next meal is coming from? Such a person is barely satisfying Maslow's most basic physiological needs. Is there more rock bottom than that? At some point, a person is just broken, and the approach of letting them suffer for another decade or so before maybe they snap out of it or maybe they die of exposure seems less "tough love" and more like just dickish neocon callousness. And, like all neocon schemes, it's expensive. I guess they could just tell hospitals to let people die in the emergency room waiting area if they can't prove that they can afford care, but I don't hear them going that far. Maybe they need to be called out on this one: does teaching someone a lesson mean that if they don't have proof of health insurance (or a fat bank account balance, or cash) on their person, they don't get admitted to the emergency room? Is that the plan? Because one way or another, our government is paying for these people. How far out do the guardrails need to be, to coddle less and let people raise themselves up out of the gutter by their bootstraps more? Put another way, what's the point of putting guardrails at the bottom of the ravine?

With the new movie out, Al Gore has been getting smeared again as a liberal wingnut. The favorite example is that he claims to have invented the internet, which is not true; he neither invented it, nor claimed to have invented it. Two of the key contributors to the technical creation of the internet clarify.

Joel Spolsky expresses in his typically entertaining and technically accurate style why powerful programming languages matter. I once worked at a company whose CTO intentionally hired smart people with a preference for fancy heavy-duty programming languages (including myself and Fiid) and made them use watered-down rapid application development tools. The idea was that smart plus E-Z = maximum productivity, as though programming languages were either easy to learn and super productive, or hard to learn and took forever to program in. Joel posits that the opposite may be true, and yet his company's flagship product uses Microsoft's rapid development language for some reason. In The Road to FogBugz 4.0: Part III he talks about how they wrote an ASP to PHP translator, so the VBScript decision there and the later bashing on Java (which C# is a clone of, and C# is a far more sophisticated langauge than VBScript) seems out of whack. I don't get it. Not that Java doesn't need bashing - I use Java quite a lot, and it needs bashing. I'm looking forward to learning Ruby soon. But bashing Java for language features and using VBScript instead? Hmm. I haven't seen VBScript for years; is it really a more powerful language than C#? Seems unlikely.

Someone else wonders why their solution to portability was to pick a non-portable technology stack and then roll their own source code translator instead of just picking a portable architecture in the first place. Good question. It appears that Joel is a great manager and a mediocre architect: anything that isn't .NET on Windows on Dell is a "religious" decision. How about C# on Mono instead of PHP, then? No need for a fancy langauge translator. That probably would have been my approach.

Speaking of issue trackers, last year I did a formal consulting-style evaluation (with weighted scores and multiple evaluation stages and all that) of issue trackers. My favorites were JIRA, Eventum, and FogBugz, in that order. Trac, which one co-worker was totally a fan of, didn't make the grade for many reasons, including crappy documentation, many lacking features, and basically no installer or migration tools whatsoever. One thing that really blew my mind was that there was no reporting or querying user interface. You just train your QA and project management folks to be database programmers, and then they'll be ready to enter raw SQL queries, which is what the Trac folks thought was a reasonable way to handle reports. Our PM vetoed it, which I agreed with... just ridiculous. Well, the secondary problem of giving end-users direct database access (I know, I know) turns out not to have been a brilliant decision. If you change the data model, which software developers need to do from time to time to support new features, all the SQL queries for reports break. Oops. So now they're taking that out and replacing it with a slightly better and more abstracted query language. Perhaps in a few more years they'll create a form based interface like everybody else?

And there's more:

Shawn: one thing I am about to add to our trac is due date custom field
Shawn: isn't that ridiculous?
Shawn: no due dates on an issue tracker?
Shawn: and that's the most popular issue tracker on the internet
Shawn: lack of due dates probably explains why trac is so popular in open source projects

One thing Joel did figure out 100% right is the software pricing issue. JIRA is bloody expensive at $1200. FogBugz is way cheaper at $129/user. For a 2 or 3 person project starting on a shoestring budget, that's huge. That's a couple of PC's. JIRA is very sophisticated, but $1200 to get one developer tracking bugs at a startup is just dumb pricing. Organizations don't switch bug trackers unless they absolutely have to, and having been in an organization that absolutely had to, I did the work to pick one (no point in migrating to another sucky tool and then having to do it yet again) and the work to migrate all of the bugs. It utterly sucks. Forcing customers to buy someone else's tool first means they'll never come back to JIRA when they have a big enough team to justify a $1200 bug tracker. Remember, there's always Trac and Bugzilla and Eventum at $0, and FogBugz at $129/seat. Atlassian is just screwing themselves with this pricing.

Funny: The Big Lebowski: The Fucking Short Version, and The Dude's Version. Mordots has its own domain now that explains the Flash interpretive video. This and the Cloudsong serious business video (loud and full of cussin') from months back make me so totally glad I'm not an MMORPG player. Better to just laugh at the freakshow.


August 8, 2006

Regarding Zune: what we know, think we know, and don't yet know: I'm excited about Zune, and that's not because I want one. I suspect that like most Microsoft 1.0 products, it will be stupid and lame, kind of like this dumb commercial which supposedly is not actually a Zune commercial but is quite boring and fails to be about anything in particular, such as a product or service or problem to be solved. I'm excited because whatever they do will force Apple to compete.

All that I've read about Zune is very much the 30,000 foot strategic pitch, which is that it'll be some devices and some software and some services. So far, it's indistinguishable from what .NET was in 2000 or so: a brand for some products which will, like, totally be awesome. Details shmetails, we'll make those up later when we decide whose products we want to emulate, I mean which products we want to innovate.

Apple has driven the wedge into the entertainment industry and into the consumer market, so Microsoft has a fair chance of being able to make deals with record companies and film studios now, provided that they implement a similar DRM scheme that annoys paying customers and barely slows down file traders. Likewise, assuming that they do something very un-Microsoft-like and support formats like Ogg and Real, all the iPod-haters will have something other than the Creative devices to flock to. They've already said they'll throw money at you in the form of buying your iTunes purchases in their own store to help you migrate, and probably the Xbox-like practice of building a device that costs much more than their competitor's device, and then subsidizing that cost so that the retail price is similar but the specs are superior.

The real hurdle will be design and marketing, in my opinion. Most of the iPod clones out there have better features from a technical checklist perspective, but are a pain in the butt to use, as though software was an afterthought, which it probably was. That's standard in consumer electronics, unfortunately: fulfill a technical feature checklist, put it in a pretty enclosure, and let the folks who picked the embedded OS and wrote the device drivers decide what the actual user experience is going to be once they turn the thing on. Genius!

Microsoft is learning (in the Xbox and Media Center teams, though clearly not in the Windows Media Player team) how to make a device that doesn't utterly suck to use. Will they get it right? Probably in a few generations, after being slammed by the media for being late and clueless. In this case, though, the generations will probably be pretty closely spaced, unlike the Xbox. They'll probably be distracted by Linux-on-Zune users, which will "steal" from Microsoft by buying devices at the below-cost retail price and then refusing to buy DRM-encumbered content to pay Microsoft back for the low hardware price, and DRM breakers, who want to transfer purchased content to unauthorized devices, such as an iPod or TiVo. They'll probably put out a bunch of goofy devices that nobody wants, like a Nintendo DS killer and a PSP killer that don't play games but instead just play Windows Media content, just because those devices exist and Microsoft won't be able to resist having an answer to everything their competitors make. If they have any sense, these devices will run Windows Mobile and be fully compatible with WM handhelds, so that they'll fill the iPod-plus-organizer niche that nobody seems to want to fill (except Palm, whose solution sucks).

But even if the design is good, Apple has gone from horrible marketing to world-class marketing with the return of Steve (and the products have made a similar transition), and the iPod marketing campaign is no exception. SanDisk's iDon't campaign was funny, but ultimately ironic and kind of sad given that their Sansa products are blatant iPod clones. (Other shapes do exist...) Microsoft marketing, on the other hand, is terrible lately. Back in the days of Word vs. WordPerfect, they wisely used feature comparison lists and direct price comparisons to make potential WordPerfect buyers feel as if they were being ripped off. The Windows 95 launch wisely made such a big to-do about something as nerdy as an OS, that it would seem like you could go see Windows 95 and maybe get a backstage pass and hook up with its groupies, a la Almost Famous. At least it would make everybody's life much more fun. It's a good ad with a clear message: everybody will use this, and it will benefit their work and play alike. Nice! And the music was a good choice: rock, but classic rock. Rebellious, but not too rebellious. Exciting and hip but not threatening, and familiar to the age group that signs the big purchase orders and to the individual home and small business folks who probably wouldn't spend money on an OS unless there was a damn good reason to do so.

But now we have the retardo Office Dinosaur campaign. Really, insulting your own customers and comparing them to slow, lumbering animals that all died out? Brilliant. Write us an enormous purchase order and undertake a massive IT project, and in return we promise to stop making fun of you, at least until next release. Who are the ad wizards who came up with that one?

Microsoft has so utterly jumped the shark. They've got lots of chances to turn things around, and an enormous war chest and installed base to help them lumber along for at least a decade, but that doesn't mean they will turn things around anytime soon. Vista is a total yawn, and much less exciting than XP was, which in turn was less exciting than Windows 2000, which was in turn less exciting than Windows 95 and NT 4 were (at least to me). I was really happy when I got to upgrade from NT 4 to Windows 2000; I only use XP because of the unavailability of security updates for Windows 2000. IE 7 is a total Firefox/Opera wannabe, superficially, still reportedly failing in the security and standards compliance departments as always. Microsoft is again a victim of its "shoddy design plus meticulous implementation quality" culture, as seen in the WMF back-door, and in the new IE 7 that is stuck being IE 6 compliant instead of standards-compliant, because they screwed it up from the start and now they're stuck supporting all those sites.

I really thought Ray Ozzie might make a dent, but so far, he's shown only that he doesn't really do blogging except once every month or two, which is a far cry from Scoble, who arguably had the best grasp on what was actually going on in consumer technology at Microsoft because he exposed his thoughts to the public and had to actually deal with their feedback. Instead we get the guy responsible for two very hyped but closed and proprietary groupware platforms that ultimately failed in favor of open internet technologies: Lotus Notes and Groove Virtual Office. It's like they were looking for a guy to help them not succeed at embracing what customers really want and what power users really are doing. It looks like they got it.

So, maybe Zune will be an example of the Xbox 360 version of Microsoft, doing something genuinely excellent after a few iterations of improvement on mediocrity. That would put pressure on Apple to build cheaper and more open iPods, and we all win. Or, they'll be the MSN version of Microsoft, and put something out there that they throw nigh-infinite heaps of money at and yet they still struggle to be in third place, gradually adding features that competitors already have had better versions of for a year or two. Even by doing that, though, they would force Apple to compete, instead of being the Microsoft of digital music that they are today. And we all win in that case, too.

So, bring on the Zune.


August 3, 2006

I rebuilt and reinstalled the OS on my home server recently. There were a bunch of small problems that were never worth fixing individually but when a fan died and the motherboard suddenly forgot how to start the booting process, I made a bunch of changes all at once.

Key improvements: Ubuntu instead of Debian (more up to date OS components), an 64-bit kernel, VMWare finally works, over a terabyte of disk space, hardware RAID (yes really, using a 3ware 8006-LP card).

The only tricky migrations were Subversion's database which was in Berkeley DB format, and the Gallery photo gallery. BDB is architecture-dependent, which means that when I tried to use it under the 64-bit OS, it wedged badly. I had to get Linux (in this case, Ubuntu 6.06) for i386 running in VMWare so that I could dump the database using BDB under a 32-bit OS (which is what the old OS was) and then reload it under the new 64-bit OS. Ugh. That didn't take long, once I realized I had to do that. The photo gallery still doesn't work and although I can identify the exact problem I don't know why it's happening and I dread the prospect of having to manyually hack the database contents to bring it back to life.

There were some problems with Dovecot also: it creates index files in .imap subdirectories (so if you have a mailbox at ~/mail/Foo, it puts the index for that mailbox file in ~/mail/.imap/Foo). Those are apparently also architecture-dependent, so the Dovecot IMAP server crashes when it tries to read them if they were made under another architecture. In my case, Thunderbird keeps trying to connect, so the IMAP server crashes in a very quick loop many many times. Eek. But it's actually easy to fix the indexes. They're built as needed from the mailbox files, so all you have to do to fix things is to delete them:

find ~/mail -type f -iwholename \*.imap\* -exec rm "{}" \;

Generally, though, I'm happy with the results - the setup is much simpler than the old software RAID + LVM setup that made it such a beast to manage, and the 64-bit OS seems to solve the VMWare problem, which means I should be able to toss the old PIII server that I've kept around to run Windows.