Smalltalk Tidbits, Industry Rants

Information on the ESUG 2006 conference tracks has been posted. Also, check here for a listing of presentation materials for past conferences. I'll be attending; see you in Prague!

It's time to look at the logs again. Traffic seemed a bit down, I think due to the short week after Memorial Day. In the meantime, BottomFeeder downloads stayed stable at a rate of 158 per day:

Platform	BottomFeeder Downloads
Update	331
Windows	296
Linux x86	120
CE ARM	99
Mac X	91
Mac 8/9	63
Linux Sparc	19
HPUX	19
Solaris	17
AIX	16
Windows98/ME	12
Sources	11
Linux PPC	10
SGI	3
ADUX	2
CE x86	1

Next, on to the HTML page accesses:

Tool	Percentage of Accesses
Mozilla	59.6
Internet Explorer	28.4
MSN Bot	3.7
Other	4
Megite	1.7
Opera	1.6
Jakarta	1

Mozilla dropped slightly, and IE edged up. I wonder if that's a one week artifact, or if more people are trying IE 7? Time will tell, I guess. On to the feeds:

Tool	Percentage of Accesses
Mozilla	25.4%
BottomFeeder	21.1%
BlogLines	8%
Net News Wire	8%
Other	7.9%
Internet Explorer	5.1%
Safari RSS	4.6%
Google Feed Fetcher	3.5%
NewsGator	2.2%
RSS Bandit	1.5%
Planet Smalltalk	1.5%
BlogSearch	1.1%
MSN Bot	1.1%
SharpReader	1%
JetBrains	1%
Feed Reader	1%
Java	1%
Jakarta	1%
News Fire	1%
RSS 2 Email	1%
Liferea	1%
Everest/Vulcan	1%

That looks about the same as always - still an awful lot of tools in use there.

Here's a hot tip for marketers who use phone surveys - my time is valuable, so if you want me to answer, you'll have to:

• Use people who speak English well. If I can't understand them, and they can't understand me - then I don't have the time to be bothered
• If your people ask a question, and I answer categorically (i.e., I say "I have one child"), then the survey shouldn't proceed with questions that ask about other children. In other words, pay attention.

If you can't be bothered with those two things, I can't be bothered with your survey.

This is interesting - the "stock" binary search algorithm fails (for the mainstream statically typed languages):

 
1:     public static int binarySearch(int[] a, int key) {


 2:         int low = 0;


 3:         int high = a.length - 1;


 4: 


 5:         while (low <= high) {


 6:             int mid = (low + high) / 2;


 7:             int midVal = a[mid];


 8:


 9:             if (midVal < key)


 10:                 low = mid + 1;


 11:             else if (midVal > key)


 12:                 high = mid - 1;


 13:             else


 14:                 return mid; // key found


 15:         }


 16:         return -(low + 1);  // key not found.


 17:     }


 

Line 6 causes problems if (low + high) overflows the size of an int. This is a non-problem in Smalltalk, of course; you seamlessly get large integers and everything "just works". I found the fixes amusing; they are all ways of coding around the limits of the type system.

Via Chris Petrilli - the shocking experiments involving diet coke and mentos :)

Michael has posted his photos from Smalltalk Solutions (and an ancillary visit to the Ottawa STUG).

Via David Weinberger, I see that truth continues to be stranger than fiction. I don't think I would have believed that a (trying to be) serious "Captain Copyright" site exists unless I saw it myself. So it looks like I'm violating their policy by linking, because of this bit from their access rights page:

"Permission is expressly granted to any person who wishes to place a link in his or her own website to www.accesscopyright.ca or any of its pages with the following exception: in order to protect the moral rights associated with this site, permission to link is explicitly withheld from any website the contents of which may, in the opinion of the Access Copyright, be damaging or cause harm to the reputation of Access Copyright. Specifically, permission to link is explicitly withheld from sites featuring pornographic, racist or homophobic content. If you link to or otherwise include www.captaincopyright.ca on your website, please let us know."

Heh - cut down, that says: "You can link to us only if you agree with us. Otherwise, not so much". I think I need to quote Bugs Bunny: "What a bunch of maroons!"

You can file this one under "why don't things like that happen when I play golf?"

Mini-Microsoft provides some links to a funny Microsoft focused Da Vinci Code riff. Scroll to the bottom; there are four five parts so far.

Perhaps the worst law passed by any government recently is the DMCA. Sure, it passed back in the late 90's, so calling it recent is something of a stretch - but the bad effects are still piling up. The DMCA didn't just regulate how copyrighted material could be used - it outlawed entire categories of real (and potential) products. A good example? The ReplayTVs we have in our living room and family room. Both offer internet sharing of content (effectively useless due to bandwidth issues) and commercial skip. The manufacturer was driven out of business by the wolf pack of lawyers sent by the MPAA - out to make sure that I watch every commercial that comes my way.

That's hardly the only example. Consider DVDs - As the IEEE says:

You're likewise out of luck if you're looking to buy software that lets you copy a DVD onto your laptop's hard drive; it's no longer for sale, at least not in the United States. Even if you want to put the movie you bought onto a pocket-size video and game console, such as Sony's PlayStation Portable, which allows users to watch video stored on flash memory or a miniature hard drive, you can't legally do so, because you'd have to “rip,” or decode, it to make the transfer—and the studios claim that this action violates the DMCA. When you rip a CD, be it to an audiotape or an MP3 file, you're not breaking any laws. But to rip a DVD you need to somehow get around the encryption technology built into a standard disc, and since such circumvention is forbidden by the DMCA, if you rip a DVD, you are breaking a law. Under the DMCA, legality doesn't depend on how the copy will be used but rather on the means by which the digital content is copied.

You have to ask yourself - why can you rip a CD and put it on an iPod, while you can't rip a DVD and put it on the video iPod? Are the two so different? The RIAA would love to make CDs the only way to listen to music, and put them under the same regime DVDs are under. Fortunately, that's probably not possible anymore - but video continues to get locked down - the "broadcast flag" might not be designed to piss off customers, but it will sure have that effect. Imagine - you set your DVR to record the latest episodes of "Dr. Who" - you don't get around to watching it for a few weeks, as other things come up - and *poof* - the DVR will have helpfully blown the content away, and not because of any settings you put in for how long to hold onto the content. Joy.

The basic problem is that copyright has been turned on its head. It's supposed to be a short term incentive that will encourage creative works. Instead, it's been turned into a long term lockbox for corporate icons (Mickey Mouse, anyone?). It's high time to drop the copyright terms back to something reasonable, and stop beating the end customer of content until morale improves.

Go ahead and read the entire IEEE piece - it's worth the time.

My last post linked to an IEEE article that illustrated the kind of technological damage that the RIAA and MPAA are (and are trying) to inflict on the rest of us, in a desperate attempt to preserve their existing business models. The good news is, the internet itself is going to make that quest very, very difficult. Consider AllofMP3, a music site out of Russia. There's an NYT story on it this morning, which points out what the RIAA sees:

Music industry officials say AllofMP3 is a large-scale commercial piracy site.

"It is totally unprecedented to have a pirate site operating so openly for so long," said Neil Turkewitz, executive vice president of the Recording Industry Association of America, which is based in Washington.

The reality is, it's legal in Russia right now, and they accept standard credit card payments - which makes it accessible to anyone with net connectivity. The RIAA can't touch it, because it's out of their reach. That's where things are headed, unless the RIAA removes their collective head from their posterior lobes. They can either let us buy music (etc) on reasonable terms, allowing us to move that music around on our own devices - or they can watch as we do it ourselves, using offshore sites in jurisdictions that don't care.

You think the RIAA will shut all of these sites down? Sure, sure - just like various governments have managed to shut down money shielding in various countries. If governments can't stop the flow of huge piles of cash outside of their tax systems, I doubt that the RIAA will be able to stop the flow of MP3s.

They really have two choices: Be part of the game, and adjust their business models, or watch as the game happens without them.

I got asked why I turn comments off on older posts - "older" meaning anything that's aged off the front page (and out of the feed). Well, the simplest way to explain that is a snippet from a log I keep of spam attempts on the server. Not a lot of detail here, but here it is:

<<< Matched Against: insurance-ratings >>>
<<June 5, 2006 10:43:45.330>>
<<SPAM: Text Match>>
<<June 5, 2006 10:43:45.455>>
<<COMMENTS OFF>>
<<June 5, 2006 10:43:45.665>>
<<COMMENTS OFF>>
<<< Matched Against: insurance-ratings >>>
<<June 5, 2006 10:43:46.428>>
<<SPAM: Text Match>>
<<< Matched Against: insurance-ratings >>>
<<June 5, 2006 10:43:49.477>>
<<SPAM: Text Match>>
<<June 5, 2006 10:44:39.131>>
<<COMMENTS OFF>>
<<June 5, 2006 10:45:05.746>>
<<COMMENTS OFF>>
<<June 5, 2006 10:46:04.345>>
<<COMMENTS OFF>>
<<June 5, 2006 10:46:08.027>>
<<COMMENTS OFF>>
<<June 5, 2006 10:47:26.502>>
<<COMMENTS OFF>>
<<June 5, 2006 10:47:44.774>>
<<COMMENTS OFF>>
<<June 5, 2006 10:54:42.368>>

The "COMMENTS OFF" bit shows that someone tried adding a comment to an older post. I have the content saved in a directory - it's all spam. The text match lines show matches against my black-list, which is a simple text file of "bad text". That catches an amazing amount of crap too (I use the same approach on the Wiki).

Anyway, this is why I turn off commenting on older posts. It would be too much for me to keep up with, adjusting the filters. I have the ability to leave comments open on individual posts, and when I do that, I get a post specific feed (which allows me to monitor it).

I also use a simple "too many hrefs" filter - I figure anything that comes in with more than a small number of links is junk (and while that has snagged a few good posts, it mostly catches link laden spam. Anyway, that and an IP throttle, and I catch most things. I'm getting tired of the trackback spam - I rarely get actual trackbacks. I may just turn those off completely.

It's a bad thing when you run out of dates.

The RIAA must be so proud:

Officer Goins stopped a car because he noticed an equipment violation. As Officer Goins exited his vehicle, McLaughlin, the driver and sole occupant, leaned toward the passenger seat. When McLaughlin told Officer Goins he had a firearm, Officer Goins observed a semi-automatic pistol on the passenger seat and several compact disks (CDs) in the car. Officer Goins suspected the CDs were "pirated," because they were in a "poor quality made CD case with the labeling." He requested assistance from two other officers that had received training concerning CDs. Minutes later, Officers Barker and Perkins arrived. Officer Barker testified that he saw CDs on the front passenger seat and on the floorboard of the car. He testified that "based on [his] training with the recording industry the thin cases and the homemade labels in the cases led [him] to believe they were bogus CDs." He explained: They were thin case CD's and the labels on them were real blurry. You couldn't really make out the reading on them that well. You could just look at them and tell that they were bogus. Concluding that the CDs were illegitimate, the officers seized the CDs they saw and searched the car for others.

During the search, illegal drugs were turned up. The troubling thing is that the search was made due to "bogus" CD's. Better not burn that podcast to a CD in order to play it in your car - your RIAA trained local police force will seize them and accuse you of piracy.

Steve Rubel has a lead on a new Google product:

The Wall Street Journal is reporting that Google tomorrow will launch a Web-based spreadsheet application on a limited test basis. No further details are available.

Hmm - I wonder if this is a "great minds think alike" thing (DabbleDB), or an attempt to jump into the slot that Avi is creating?

Steven Dekorte:

So if we write generic code that actually takes a "type of anything," that type can only be an Object, and our generic code must only call Object methods on it. So really, we are restricted to code that is already "generic to Object," except for casting up to Object and down from Object, which this wonderful new syntax will do for us. Sounds like it's a solution for collection classes and not much else, doesn't it?

As opposed to languages like Smalltalk (Pythin, Ruby, etc) - where any method you write is able to deal with any object that conforms to the API. As Steven puts it:

Languages like Python, Ruby Smalltalk and Io on the other hand, compile a single method which will work at runtime on all types which respond to the talk method - even ones defined at runtime and all without defining the types of the variables in the calling methods.

You can have simplicity, or complexity. Which makes for a more productive developer?

More evidence that the US PTO needs a severe beating:

Internet phone company Net2Phone has filed a patent infringement lawsuit against rival Skype Technologies and its parent company, eBay.

...

Net2Phone alleges that Skype, a wholly owned subsidiary of eBay, violated its "point-to-point Internet Protocol" patent. The patent calls for the exchange of IP addresses between processing units in order to establish a direct communications link between the devices via the Internet.

A patent on point to point IP? You mean, a patent on some of basic plumbing of the internet, or do they think that the form of the packets they send between hosts are somehow innovative? As to p2p, it's not as if that hasn't been done before, either.

TechDirt notes that there are an increasing number of artists who are embracing the web - and not buying into the RIAA's chicken little rhetoric. Which begs the question: other than the labels, who exactly is the RIAA helping? If the artists are starting to wonder, then the rest of us should as well.

My cousin is a part time musician, and he's got nothing but contempt for the RIAA and its actions. None of it is helping him or the band he's part of.

Steve Rubel notes that Leno is available on iTunes (has been for awhile, I guess - I just hadn't noticed). I suppose the RIAA's goons would prefer that he sell this stuff months later on CD, but - as I noted in my last post - the artists (and apparently, some of the content owners, like NBC) understand the changing landscape:

NBC tomorrow will expand The Tonight Show with Jay Leno presence in the iTunes store, which to date has received lackluster reviews for its lack of depth. According to the network's PR rep, NBC will make available Leno's full monologue and comedy sketches the very next day on iTunes for $1.99 each. A “multi-pass” package pulls together a month’s worth 20 downloads for $9.99.

I expect to see a lot more of this.

Learning Seaside points to www.loop.aero, which seems to be using Seaside.

Chris Petrilli links to Steve Dekorte, who has a post up on the idea. Here's Chris:

First, pre-fab offers the ability for substantially higher quality materials at a lower cost. For example, by assembling pieces of a house in a controlled atmosphere (i.e. a factory v. rainy outdoors), you can control the bonding of various elements. In addition, by using larger tools and jigs to cut and form things, you can make sure that the things are repeatable.

Second, by manufacturing in larger volume at once, you can make sure that things are “right” the first time. Most problems I’ve seen in construction stem from the ad-hoc nature of construction. Even when it’s two identical floor-plans, the walls are in slightly (sometimes as much as 6-8”) different locations.

While I agree in general, it's not going to be easy to move off the ground we are on now. In my neighborhood, for instance, most of the builders were "custom" builders. Meaning, they had floor plans, but gave us (the homeowners) a bunch of options - including ad-hoc ones we thought of - to customize. We ended up changing a lot of things, and I know a lot of other people in the community did as well. This is a model that people understand, and getting them to shift will be hard work.

Now, there's nothing that says you couldn't get better results with a "mass customization" pre-fab style of building - heck, you might end up with better results (in one room where we wanted a cathedral ceiling, we got a tray instead, due to construction issues). The hard part will be getting buy-in from potential homeowners - and getting past the existing builders, who will likely resiist this idea every bit as much as the music labels resist change.

Suzanne and I spoke to James Governor and his colleague Michael Cote' at RedMonk - it was a good conversation. They asked us a number of good questions and gave us some things to chew over as well - it's nice to connect a voices their blogs. We'll certainly stay in touch as we take our products forward.

Oh, and something interesting that James and I agreed on early in the call - James McGovern had a good post up this morning that I agreed with (will wonders never cease :) ).

I guess this qualifies as "if I can't see it happening, it's not work" - HP is moving some of their IT staff away from telecommuting and back to office life:

The architect of the HP division's change, Randy Mott, is regarded by Wall Street as a mastermind of operational efficiency based on his days as chief information officer at Wal-Mart Stores and Dell. Since joining HP as CIO in July, Mott's philosophy on building a strong IT workforce starkly contrasts with that of competitors, who encourage telecommuting to retain skilled workers who desire better work/life balance.

Mott said by bringing IT employees together to work as teams in offices, the less-experienced employees who aren't performing well -- which there are "a lot of" -- can learn how to work more effectively.

In an office, "you're able to put teams together that can learn very aggressively and rapidly from each other,'' he said.

The problem is right there in the second paragraph. If there are "a lot" of ineffective employees, then you have a problem that won't be solved by putting them in an office. What you have is a management problem.

Now, not everyone is capable of working out of a home office - some people really do need the close, daily interaction with other staff in order to be effective. I rather expect that this move will lower morale in the group in question though, and will lead to the higher performing staff jumping ship. I think it would have worked out better to identify the ineffectual employees and replace them. That would have required actual management effort though, which this guy clearly didn't want to expend.

Scoble on the Google Spreadsheet:

Ahh, middleware 2.0 wars coming soon to a browser near you. Why? Cause as Google gets more people to try its spreadsheets more people will ask for more features. If they don't get those features the PR will turn back toward Microsoft's approach (since our Office has a lot more features than Google's offerings do). There will be pressure on at Google to add features but DHTML (er, Ajax) will simply run out of gas. So, you'll start seeing middleware coming down. (Runtimes like .NET, Flash, Java, and WPF, are what I'm thinking about -- I'd bet that Google is working on a browser-runtime of its own that'll add a lot of local functionality to Web clients).

I don't think most users of this kind of software will want lots of new features. The target for this isn't power spreadsheet users - it's small groups that want to collaborate with (relatively small) datasets. We've seen an example in this space already: DabbleDB.

Consider most spreadsheets you see. Do most people use any of the non-basic features? Or do most people do a quick set of calculations, and possibly some simple charts? I'd say it's the latter, by a very wide margin. I think Robert misreads the target audience for this kind of application, and even for the spreadsheet in general. Heck, I'm in Product Management, and I can tell you that I create only very simple spreadsheets - which is what I suspect most people do.

Sure, there are power users, and they use a lot of the advanced stuff. I doubt that any of those users are serious targets of this kind of online application.

The LA STUG is meeting on June 12th:

LASTUG Meeting

Monday June 12, 2006
7:00 pm - 9:00 pm

This event repeats on the second Monday of every month.

Event Location: High Tech High, Los Angeles - Meeting Room
Street: 17111 Victory Blvd City, State, Zip: Lake Balboa, CA, 91406
Map

Notes:
There is usually an after meeting meeting at Jerry's Deli in Van Nuys that goes on to an indeterminate time.

If there is a problem getting there call Darius Clarke, Mike Klein or John Dougan for assistance. The phone numbers are in the lastug contacts database on Yahoo!

The NY STUG is having a joint meeting with Patrick May, of the NY Ruby user group tomorrow evening. Details on that are on Charles' blog.

If you like Smalltalk - any implementation - do you want to go where we are going with Cincom Smalltalk? Or where these guys are going - which is apparently deeper into Eclipse-Land. Door number one includes a lot of improvements and advancements in Smalltalk, while door number two is focused elsewhere.

I saw the news that Apple was pulling out of India the other day, and - since it was call center/support related, chalked it up as another example of management actually hearing the tooth grinding of their customers. Apparently, I was wrong, as Paul McDougall of InformationWeek has analyzed it and come to a different answer - he relates it to IBM's expansion there:

IBM is betting big that India will for years to come be the center of excellence for software development and related work like IT services and help desk support. With this in mind, IBM wants to flat out own India's technology landscape and it's apparently CEO Sam Palmisano's belief that half measures won't do. There's too much potential competition down there on the Subcontinent.

And Apple?

Apple's decision to shut down its services center in India just three months after its opening is the right one only in so far as the company really had no other choice. The Mac and iPod maker realized it's just too late to the party in India. There are now too many companies chasing Indian IT talent to try and build a base there from scratch. In all likelihood, IBM's plans to spend billions more in the country put the final nail in Apple's own Indian ambitions.

Hmm. So India is full, what with all those IBM suits running about - there's no way Apple can get any kind of operation running there, so they may as well give up. Well. It's really too bad that Google never happened here in the US, what with the IT sector being full and all - the talent was elsewhere, so Google just dried up and blew away, right Paul?

Perhaps Not.

There's a phrase Paul might want to examine: "Sometimes a cigar is just a cigar". Perhaps Apple's pullout has no inherent relationship to IBM's actions there. I seriously doubt that anything like this happened in Cupertino:

Jobs: So why can't we hire more smart folks in India?
Staffer: Sorry Steve, IBM hired them all. That <expletive deleted> Palmisano at IBM hired every last IT guy on the subcontinent

Jobs: <expletive deleted>! Just pull out then, it's all we can do
Staffer: Yes sir!

That seems to be how McDougall sees it. All I can think of to say about that? *LOL*

Jon Udell sounds like he'd like to offer multiple enclosures (given various format needs), but - there's a problem. The *cough* RSS Spec *cough*:

So, to Greg's question, which of these formats should be in the enclosure? Flash, QuickTime, or WinMedia? (Additional wrinkle: the Flash FLV format requires delivery of a Flash-based player, as well.) If I enclose all three, that starts to get hefty. And last I heard, sticking multiple enclosures into an RSS 2.0 item is a nonstarter

Of course, Winer could have written that into the spec. Heck, he could stop being a pain in the neck, and let someone else add it to the spec. Then again, monkeys might fly out of my butt, too.

If you're my age, this collection of '80's videos will bring back a few memories. Who knows if the content is online legally - watch it while you can.

Hat tip James Governor.

Dave Winer throws another tantrum:

How can we be friends if he's friends with people who are selfishly trying to monetize our work, without giving anything back. I say "our" work, because Toni represents many thousands of WordPress users. There's an arrogance around Battelle's conference, they're the insiders and we're the poor schnooks whose work they monetize. I'm one of those people. Anyway, I think this is beginning of a valuable discussion. Perhaps we can help Battelle and O'Reilly straighten this out, if that's what they want. If they don't want to, then we shouldn't be supporting their conference.

Translation: "Wahhhhhhh - they won't invite me to their kooool conference!"

Tim Bray has a good post up on binary search, but prefaces it with this:

Anyone who regards themselves as a serious programmer has internalized a lot of different ways of searching: hash tables, binary, and many different kinds of trees. I've used pretty well all of these seriously at some point, but for a decade or so, as far as can I recall I've used almost exclusively binary search, and I see no reason to change that. Herewith an essay for programmers, with fully-worked out examples in Java, on why. [Updated 39 months after publishing when I read with horror Josh Bloch’s exposé of a long-lurking bug. If we can’t get binary search right, what chance do we have with real software?]

As I said here, the code is fine. That type system you're using with Java? Not so much. Patrick Logan makes a related point here. The chances of getting software right will improve a whole heck of lot if we just stop using broken tools.

Yet another media guy who's offended at the idea that us paeons out here might be creating content:

The video-sharing site YouTube is a poster child for this sensibility, since anyone can upload just about anything to it. For a sense of what this new world is like, you can consult the site's "Top Favorites." There are several dance segments, people imitating ninjas or lip-synching songs, and a (very funny, actually) dog who growls at his own leg. You can spend 10 minutes and take in all of it. Spend much more, and you start feeling guilty about the time you're wasting.

That's right boys and girsl, it's all crap, because no professionals were involved. He also casts a dim eye at blogs:

Another way that people describe mash-ups is "user-generated content," referred to by the smart set as "UGC." Most of the time, when companies talk about user-generated content, they mean nothing grander than the pictures you store on Web sites or the pages that MySpace members spend hours fussing over.

Talk about framing. Hey Lee - 10 minutes spent with "The Enquirer" is mildly amusing, while more than that feels like wasted time. Can I extend that argument to your scribblings too? After all, you're paid media, and it's all the same thing, right? So what's worthwhile?

These aren't all tweedy costume dramas. No. 1 is "Fawlty Towers." No. 2 is "Cathy Come Home," a Ken Loach drama about the homeless that first aired in 1966 but is still vividly remembered. The rest of the list includes dramas and sci-fi and talk shows and sitcoms, all of them, in their own way, weighty meals for the mind. You can watch them decade after decade, and never feel guilty at all.

I'm not going to say anything bad about that stuff - I love "Fawlty Towers", for example. However, just because professionals create some good stuff doesn't make it all good ("Gigli", anyone?). Likewise, just because there are insipid instances of amateur content doesn't mean it's all crap. You might think a pro would get that, but hey - he's only a professional writer...

Tim Bray doesn't like the "it just works" argument:

“The poor boy, that primitive Java stuff broke because he doesn’t have auto-magical big numbers like Lisp-n-Smalltalk had back in the day.” Thank you for raising my consciousness. If you’ll grant that the trade-off between fixed-size hard-wired datatypes and more abstract ones has been under discussion since Turing was a tot, I’ll grant that many attempts to pack the data in tight are symptoms of premature optimization. But space-vs-time trade-offs are just not gonna go away; deal with it. And I’ve had my working set blown to hell more than once trying to build the parse tree for what seemed like a moderately-sized incoming message, in a language that turned out to be just a little too high level. And the “My thought-experiment language solved that in 1976” mantra is boring .

Here's another one for him: Try doing the factorial of 1000 with Java integers. Whoops - can't do that either. It's not that the space vs. time is going completely away, but: in a world where we have 1GB+ of memory available, and hundreds of GB of disk, it's an affectation to hold onto 32 bit integers as some kind of rational optimization. Face it, Tim - Smalltalk and Lisp got this one right a long, long time ago, and James Gosling still hasn't wrapped his head around it.

Ars Technica has some good news: Comcast is going to be testing the Tivo interface on their DVRs. Not a moment too soon; the existing interface really, really sucks eggs (ask my wife - she really gets exercised about it :) ) Anyway - here's the important part:

After more than a year, we may finally start to see the fruits of cooperation between TiVo and Comcast. In March of 2005, the two companies agreed to work together on implementing TiVo's interface and functionality on Comcast's own DVR boxes. Relations between the two companies had previously been complicated and strained, and in the early part of 2005 there were many fears that TiVo was ultimately doomed. Those fears haven't entirely subsided, but things are certainly looking up.

Really good news.

Update: This seems to apply to a specific line of Comcast DVRs - and mine isn't one of them. Sigh

The Mini is back, with a new hard drive. With any luck, that will solve the problems I've been having. I had backed up the stuff that was on the iPod last time, so I got my entire music library shifted back, simply by slamming it back from the external HD. Now it's busy downloading my daughter's purchased music from iTunes - it turns out that Apple does have the capability built into iTunes. That makes her happy, and I'll set up cron jobs to keep everything safe.

Hey Tim - about this last line of your post on numerics in Java:

And the “My thought-experiment language solved that in 1976” mantra is boring.

Hmm. So Smalltalk is a "thought experiment" languages, eh? While it don't have the adoption level of Java, there are two things you ought to keep in mind:

• It's used in the real world (and the vendor behind the main one, Cincom Smalltalk, actually turns a profit on it. Unlike some language vendors I might mention).
• Having a type system that is both consistent and actually works isn't a "thought experiment". How Java measures up in that regard is an exercise best left to the reader.

People just keep figuring out that things we did in Smalltalk a long time ago have merit:

SecondLife is using Mono in a non-conventional way, which I like to think is one of the benefits of having an open source engine, they have added a micro-threading implementation.

Microthreading was necessary because some of their simulations are made up of thousand of threads/routines, and using the default threads in Mono (which are mapped to operating system threads) would have been too heavy.

That's why BottomFeeder can spawn a thread per feed, even when subscribed to hundreds of feeds - because they don't map down to OS threads.

This news about the Dixie Chicks should make marketing and PR folks - and blog triumphalists - look up and take notice. Their new album has been selling very well online, but they are having difficulty filling venues for concerts:

Initial ticket sales for the Dixie Chicks' upcoming tour are far below expectations and several dates will likely be canceled or postoned.

Ticket counts for the 20-plus arena shows that went on sale last weekend were averaging 5,000-6,000 per show in major markets and less in secondaries, according to sources contacted by Billboard. Venue capacities on the tour generally top 15,000.

In contrast, the band's new album, "Taking the Long Way," sold 526,000 units in its first week, according to Nielsen Soundscan, the third-largest sales week of 2006. The album logged a second week in the period ended June 4, according to sales data issued Wednesday.

One of the interesting aspects of the net is the "long tail". Regardless of what hobby or profession you are in, the net - especially the blogosphere - makes it easy to find like minded individuals and form a community. That community might be quite large, and as with the album sales cited above, be commercially successful. That doesn't necessarily mean that you've got a mass audience in the classic mass marketing sense of the word, however.

What the Dixie Chicks are learning about first hand is the existence of the long tail, and what it does and doesn't mean. Here's a more personal example. On a weekly basis, I have about 20,000 readers. Does that mean I can promote a Smalltalk conference and expect 20,000 attendees? Based on the attendance in Toronto (which was good, but in the hundreds, not thousands), clearly not. My readers are in the long tail. Not all of them are Smalltalkers, and, of the ones who are, not all of them will go to a conference (for a variety of reasons).

The net makes it easy to mistake a large online community for a similarly large offline community. The two aren't the same. Online, geographic space is irrelevant. Offline, it's not. That has relevance for artists, marketers, and politicians, just to state three obvious examples. For the next little while, I expect to see a number of marketing errors based on this.

I've just added NTLM authentication support to BottomFeeder, but I can't take credit for it - that goes to our engineering team, Tamara Kogan in particular. In Bf, we use the NetResources package for HTTP support, and we had previously implemented full support for Basic and Digest authentication. I'm able to bypass that now, and just go to the base library support - which is great, since it's now code that I don't need to maintain.

It's not shipping yet - I have to migrate BottomFeeder to VW 7.4.1, which is not quite out yet - that's imminent. That means that the current dev version, base don 7.4, will not be promoted on 7.4. I should have a new release out shortly after 7.4.1 ships, as I'd like to get this new support out. Stay tuned.

Update: In response to a comment from Rich - what does this mean for end users? It means that when the next rev of BottomFeeder comes out, it will work with a wider array of proxy servers than it does now. Specifically, if you have an MS specific setup, you probably have NTLM (which is sort of like Digest, but non-standard). Right now, BottomFeeder doesn't work with NTLM proxy servers - meaning, you can't subscribe to anything that requires proxy services. With the next release, that problem is gone.

Another feature bites the dust in Vista: p2p file synching:

"While PC-to-PC Sync is a great feature that improves productivity and collaboration we don't have it at the quality level our customers demand," a company spokesperson said in an e-mail. "As a result the decision was made to remove it from Windows Vista."

So what's actually new in Vista? Well, the security model (which sounds really irksome, based on a few things I've read). Oh, yes - PVP-OPM is there too. Features customers might actually want? No time. Stuff the goons at the RIAA and MPAA insist on? Bring it on.

Vista is "all about something", but it's certainly not end users...

I commented on the "IBM bets on India" story the other day in the context of Apple's pullout - it occurred to me that I should look at it on its own. The gist of the story: IBM is on a hiring spree in India:

IBM said Tuesday that it will triple its current level of investment in India over the next three years, bringing its total spending in the country during the period to $6 billion. The plan aims to vastly increase the range of IBM's offshore computer services offerings. Those services are designed to help businesses cut costs, but critics say they also threaten U.S. tech jobs.

...

IBM currently employs 43,000 workers in India, up from 23,000 just one year ago. At the same time, the Armonk, N.Y.-based company has been quietly trimming payrolls in the U.S., where its staff complement is now less than 150,000. IBM officials were not immediately available to comment on how plans for India would impact the U.S. operations. On a Web site operated by current and former IBM staffers, www.allianceIBM.org, posters routinely share news about layoffs at IBM sites around the country.

They hired an eye popping 20,000 employees in India last year, which tells me something that the business analysts are missing: they are growing staff way, way too fast. I don't care whether you are hiring in your own backyard or on another continent - there's simply no way to manage that hiring pace effectively. Palmisano can boast all he wants:

"If you are not here in India, making the right investments and finding and developing the best employees and business partners, then you won't be able to combine the skills and expertise here with skills and expertise from around the world, in ways that can help our clients be successful," said Palmisano. "I'm here today to say that IBM is not going to miss this opportunity."

But he's making a huge error - and I don't mean in terms of where he's hiring (although, readers of this blog will recall that I don't have confidence in that, either). He's simply hiring too many people too fast, and he's going to end up with the same kind of hangover that the dot-bomb firms experienced in 2001. I wonder how quotable he'll be then.

The other question I'd have is on how they are planning on managing development. Here's their stated goal:

The announcement comes on the heels of a plan IBM unveiled in March, under which the company is moving all development of business solutions based on service-oriented architectures to Bangalore.

That may or may not work well. The big question I'd ask is this: Where is project and product management located? If the answer for either one is "in the US", then I'd guess that the problems are coming down the pike. It's a 9 hour gap between the US east coast and India, and 12 hours from the US west coast. That make communication very hard, because someone has to be willing to work during non-business hours - on an ongoing basis. That's not a scalable solution. If you want offshoring to work, you have to be willing to place the management right there with the developers, in my opinion.

I have some experience in this area - we have a geographically spread team. Most of it is in North America, but we have some support staff in India, and a few developers and partners in Germany. The 9 hour time gap between the west coast and Germany makes it hard to manage communication - there are short windows in which it can occur. We only have a few people that far off - if the entire staff were 9 or 12 hours away from me, my life would be hell. Bottom line, if IBM hasn't accounted for that, then they have some rough sailing ahead.

Spotted in inessential.com

I’ve been getting a ton of random-letter comment spam lately. Does anybody know what the purpose of this is? Or have any theories? The comments are weird because they’re apparently useless. No links, no words. They’re not selling anything or trying to get a better Google rank. They’re Zen comments. What’s the point of them?

I'm seeing that too, but mostly on the Wiki. The only guess I have - and it's a guess - is that someone is testing a new spambot?

Looks like there are versions of the RIAA overseas that are every bit as stupid as what we have here. Witness this gem of a story from Italy:

Internet firm Tiscali has suspended its music sharing Juke Box and accused the European recording industry of being "virtually impossible to work with".

And how is the music industry impossible to work with, you ask?

It took the move after it was told to remove the service's search by artist.

That certainly generated a "wtf?" reaction from me, so I read further down. Eventually, I reached the *cough* rationale *cough* given:

But the International Federation of the Phonographic Industry (IFPI) said Juke Box had offered a level of interactvity that breached its licence.

...

However the IFPI decided Tiscali "was paying to offer one type of service but was actually offering another very different one".

"Consumers were allowed a high degree of interactivity that breached these rules in many ways - for example, streaming individual tracks on demand," it said.

Individual streaming is "too high a level of interactivity?" - sheesh, darn those consumers for wanting to to find their favorite artists - they might want to *gasp* buy something! I have to say, when you make the RIAA look reasonable, you have a problem...

Smalltalk Spreads

It's looking like Seaside might be be bundled on the 100$ laptop being developed by the One Laptop Per Child project, as I had imagined last year.

Ethan Zuckerman reports that the latest prototype will include three development environments: Python, Javascript and Logowiki . Well, Logowiki is written by Luke Andrews, Avi Bryant, Andrew Catton, Alan Kay and Colin Putney and runs on top of Lukas Renggli's Pier ...that runs on top of Squeak Seaside.

That's pretty cool.

I guess the blog search engine shakeout has started: Mike Arrington is reporting that PubSub is going down:

Blog search engine PubSub had massive layoffs today after last minute merger discussions with knownow fell apart. It looks like a shutdown is imminent.

Google survives as a search engine via the ad business. I subscribe to a bunch of PubSub generated feeds, and I've had no reason to revisit their website since I set those feeds up. The feed items they find don't come with ads, nor do they slip ads into the feed separately. When I browse an item they found, I go to the item - not to a summary on their site.

Push all that together and it means that there's no revenue. If you give away a service, you need to find some other way to monetize it - and PubSub just hasn't.

ArcterJournal looks at Vista:

Default ram used after boot with not doing anything is 553mb. This went down to 478 when I got closed the welcome center, exited the sidebar and clicked through a couple of other random dialogs that popped up when I clicked on the RSS feed on the sidebar earlier, telling me about IE7 phishing philter.

My take away from that - if you decide to run Vista, you'll want at least a Gig of RAM. Probably 2, actually.

We are getting closer to release - we did a new build yesterday, and will be having a call on the status for release this afternoon. We'll probably let the release get a looking over by the vw-dev community for a week, but it's imminent now. Have a look here for some details. If that list looks short, recall that our summer releases are maintenance ones. Look here for what's coming in the winter.

Sometimes, I outclever myself. I added a simple text filter for screening comments and trackbacks recently. It's a simple filter, and that's part of what bit me today. I had someone ask me why their comment got eaten, so I went to have a look. As it happens, it was chewed up by a stupid entry in my filter. I added "anal" to my filter, due to a bunch of pr0n stuff showing up.

Sadly, I didn't limit that to looking for "anal" with a space after it (and "anal" could be valid, as in "that's a really anal way of looking at it"). Anyway, the comment in question got eaten by the word "analysis". Dumb, dumb, dumb.

It's fixed now, until the next dumb mistake like this I make :)

Time for the end of the week wrap on the logs. BottomFeeder downloads stayed strong, at a rate of 172 a day. The details:

Platform	BottomFeeder Downloads
Windows	339
Update	215
Mac 8/9	183
Linux x86	142
Mac X	95
CE ARM	90
Solaris	30
HPUX	29
Windows98/ME	26
Sources	22
AIX	18
Linux Sparc	11
Linux PPC	3
SGI	2

Which takes me to the HTML page accesses:

Tool	Percentage of Accesses
Mozilla	59.7%
Internet Explorer	29.6%
Other	4.5%
MSN Bot	2.9%
Megite	1.7%
Opera	1.6%

I see that the IE stats are creeping up - Surprisingly, only 2% of the IE usage is IE7, so that IE number isn't growing much due to that. The good news is, there's real competition in the browser space now. The bad news is, MS still hasn't supported CSS properly. Finally, on to the RSS accesses:

Tool	Percentage of Accesses
Mozilla	23.4%
BottomFeeder	21.6%
Net News Wire	9%
Other	9.6%
BlogLines	7.9%
Internet Explorer	5.4%
Safari RSS	4%
NewsGator	2.6%
Google Feed Fetcher	2.5%
SharpReader	1.6%
RSS Bandit	1.6%
Planet Smalltalk	1.5%
Feed Reader	1.3%
JetBrains	1%
Jakarta	1%
BlogSearch	1%
Java	1%
RSS 2 Email	1%
News Fire	1%
Lilina	1%
Liferea	1%

Mostly a normal distribution, although BottomFeeder access is a bit up. We'll see if that sticks next week, or if it's just a weekly anomaly.

The National Hurrican Center reports that we have Tropical Depression One on the map. If their maps are right, they have it hitting Northwest Florida next week. I'm heading to Florida next Thursday, but it looks like this storm will be long gone from Florida by then.

Could Duke Nukem Forever actually be coming out? Or do we file this under "Godot"?

Doc Searls notes the rejection of "net neutrality" legislation, and ponders what's next:

Now that the NN bus has crashed, maybe we can get together and think of better strategies - and not just political ones - to build the Net we want, while preserving the best of what we already have.

I'm not that worried. At the consumer side, there's already a set of tiers, depending on what you are willing to pay for. In my area, there's everything from dialup to 30mbps down, 5 mbps up FIOS. prices range from $15/month on the low end, up to $180/month on the high end. This is far more choice than I had just a few years ago, btw, and it's all coming via the dreaded carriers.

It's also tiered service - not everyone has the same internet experience, which is a lot like everything else. Not everyone enjoys HD TV, either. The reality is, things are improving in the direction we want, without some overriding governmental control. Heck, a few years back, the corporate grade connection into our engineering office in California was a T1 - which offered symmetric 1.5 mbps. I can now buy better than that for my own use.

I'm not really worried about a one way set of tracks being erected - that's not the direction things have been going, and I seriously doubt that they'll start going that way.

Update: Doc updated his post to reflect his (lack of) choices where he lives. The thing is, internet service is no different than any other product - you get more and better choices in some areas, and fewer in others. Internet service just isn't going to be magically universal and better than other things.

Yann Monclair sends word of a Smalltalk meeting in Paris, on July 18th:

Eric Winger offered (on the squeak-fr mailing list) to present Gemstone's Smalltalk products (GemBuilder, Gemstone/S 64...) in Paris.

It's been scheduled for Tuesday, the 18th of July from 7pm to 9pm.

The meeting will be held in the offices of OCTO Technology, 50 avenue des Champs Elysees, Paris.

If you are interested in attending this presentation, you can add your name on the wiki page

Sounds like fun. And remember, there's a Smalltalk party in Cagliari, Italy on July 1st

If this is true, then Sony will need to have a lineup of new games for the PS3 - this is from the guy behind the "Final fantasy" franchise:

The Xbox 360 operating system shares enough similarities with Windows, he said, so that porting the Windows version of FFXI to the 360 was a fairly quick task. A PS3 version of FFXI, on the other hand, would require redeveloping the game almost from scratch, a process that Tanaka estimated would take two or three years.

As a result, FFXI will emphatically not be a launch title for the PS3. In fact, Tanaka did not commit to bringing out FFXI for the PS3 at any time. He feels that the resources required to port the game to the PS3 might be better invested in a new game that's built from the ground up for next-gen hardware--but his team has yet to make a final decision one way or another.

I don't follow the development side of the game industry that closely - is it that the game engine for the PS3 is that different?

Michael van der Gulik posted some code that runs a number of processes, and wondered about the results - which showed that the processes don't all run:

Either:

- My code is borked, which is entirely possible,
- The ProcessScheduler is buggy, or
- Squeak is meant to work like this.

Which of those is true?

I haven't examined the scheduler in Squeak, but in VW (which is descended from the same original codebase), processes are cooperative - i.e., a process at priority N will never interrupt another process at priority N. So in VW, when I tried his test, only one of the processes ever ran. He set up N processes, and had them fork like this:

loop: element
	[[ continue ] whileTrue: [ 
		counts at: element put: ((counts at: element) + 1).
	] ] forkAt: 10.

So if I do that, only the first one in ever runs (as they never get blocked on i/o). To make them all run, you do something like this:

loop: element
	[[ continue ] whileTrue: [ 
		counts at: element put: ((counts at: element) + 1).
		Processor activeProcess yield.
	] ] forkAt: 10.

That #yield puts the process in question on hold, allowing others to run. This is simply the way VW (and, to a large extent, it seems) Squeak work. If you want pre-emptive scheduling, it's easy enough to do - just change the scheduler (all the code is there in Smalltalk).