Smalltalk Tidbits, Industry Rants

I finished Niall Ferguson's "The Pity of War" today. It's a fascinating read, but also a difficult read - Ferguson packs a lot of statistical evidence into his book. For instance, he makes the case that the Entente Powers fought a far less effective war than the Central Powers did - they had bigger economies, larger armies, and they had invested more money (both in percentage of GDP and in raw cash terms) towards their militaries. And yet - Germany nearly won in 1914, and nearly did so again in 1918. In 1914, it was the arrival of the BEF which stopped them on the Western Front - and in 1918 it was the arrival of the AEF that did it again.

Ferguson argues that it would have been better had Britain stayed out, and at the moment (I intend to read more widely on this), I find it hard to argue. Witness what we have in Europe now - a mostly German led EU. What would we have had if Germany had won? The same thing, only with two crucial differences:

• The emergence of the EU 80 years early
• An unexhausted, still powerful British Empire to check the emerging EU

Ferguson also points out that we likely would have avoided WWII, and may well have avoided the founding of the USSR. It's impossible to tell now, of course. I had an additional thought - regardless of what you think of the current state of affairs in the middle east, the configuration of that region was set into its present form as a result of the Entente victory. What would have become of it instead is hard to say, but I find it difficult to imagine a worse result.

In any case, I highly recommend this book. Whether you end up agreeing with Ferguson or not, this book will make you think.

One more nifty anti-spam trick down the drain - Troy pointed me to this lovely post:

There are some things that machines are better at doing than people, and vice versa. Automation is all about the former and CAPTCHAs - those little mangled-text images that you have to type in before you're allowed a free email account - are all about the latter.

The purpose of CAPTCHAs is to foil automated attempts by spammers to harvest tons of free email accounts. The trouble is that, as was identified over a year ago, you can automate circumvention, if you're clever about how you harness and use human processing power. In this case, you set up a site with content that people really want to get. (Porn, or warez, or... you get the idea.) In order for people to get to the content, they have to go through a CAPTCHA test - except that the CAPTCHA is actually grabbed from the web service whose defenses you want to breach. Your eager porn-surfing visitors are doing all the hard work for you.

Apparently, this isn't new news - Jon Udell wrote about this awhile back, (Yoz's article is from December - I must have missed all this good news). Of course, there's another problem with captcha's - they create an accessibility problem - which can be a legal problem, an ethical problem, or both. Just when you think the commons can't take another head shot, you find out just how wrong you are.

It looks like BEA is after the great white whale of programmer free development:

Facing slowing sales to its traditional customers, BEA Systems is trying a new route: pitching its software to nontechnical businesspeople frustrated by the slow pace of IT change.

BEA plans to introduce a new product line that will be sold under a separate brand and released in stages over the next several months, BEA executives told CNET News.com. The software is designed to let businesspeople create and make changes to Java code without the need for programmers.

BEA sells server software and development tools for building and running business applications. Its usual customers are Java software programmers and higher-level technology executives such as software architects and chief technology officers.

The company's new product line will target businesspeople, such as a purchasing director or a business analyst, who don't have technical training but understand how new software should work. BEA's goal is to expand its sales beyond the limited pool of technical professionals and Java programmers, executives said.

Heh. Yeah, sure. I believe that like I believe in fairies. Eventually, someone will have to maintain the results of the generated code. I suspect that'll be the job from hell. Hat tip Jason

I abused Gartner over the 20 year prediction crap they were pushing awhile back - just today, I was sent a link to this post by Bob Lewis (Alan sent me the link):

Perception, of course, is reality, so you need to pay attention to this, especially since Gartner sells the perception to the executives with whom you work, also sells its solution, and, unlike you, has a paid sales force and PR machine. According to the article (and to be fair, it's possible Gartner's actual findings had fewer logical holes than the article that presented them) what's going to happen is:

New technologies, that deliver pre-packaged workflows to businesses, and let businesspeople reconnect the process flows by manipulating visual tools and pushing a button (Ta-Da!!!) will fundamentally change the responsibilities of IT departments.

Since today most IT organizations spend the bulk of their budgets on operations and applications, something fundamental needs to change. Mix in outsourcing and the end of IT is at hand.

Oh, and the career solution for technical professionals? Get an MBA.

Bob goes on to deconstruct this nonsense. I know Gartner has a great reputation, and that people pay real money to hear them blather. What you should ask yourself is why? Do they predict, or do they just engage in a jolly game of making crap up? I might as well ask my daughter what IT will look like in 20 years - at least I'll get an honest "how would I know" answer out of her.

I thought I'd seen this kind of referer spam show up on my site, and I was right. Blogspot is mostly an afterthought for Google now, and the spammers are off exploting the afterthought. Here's what's going on:

A friend recently returned from a anti-spam conference and said someone gave a demo of a spamming tool. They showed how it grabbed a zillion email addresses from a database, started churning out the email while hopping from one free open proxy server to another, and one curious last step was to automatically create a new blogger account, create a new site on blogspot, and load the email text from the spam as an entry. The last step was to raise the search engine position for the spammer's site and message and was completely automated.

It's time Google did something - maybe a captcha based thing on the blog setup form? Right now, it's just an open sewer creating more sewage for the rest of us to deal with.

Scoble is right - you shouldn't assume that your boss isn't going to see your blog:

Can my employer fire me if I blog from home on my own time? Yes. The odds of your company perusing your blog is slim. "But if your boss should see your blog and be offended by something there, in most states you have virtually no protection against being fired," says Lewis Maltby, president of the National Workrights Institute in Princeton, N.J.

To which Scoble answers:

I disagree with something on the first page of that article. The odds of your boss reading your blog are NOT slim anymore. More and more bosses are figuring out that by using an RSS News Aggregator and Pubsub.com that you can keep up to date on what ANYONE says about your company on any of the eight million blogs. In fact, this is so efficient that usually I see new things within minutes of them being written now.

Very, very true. I run across comments made about Smalltalk, Cincom, or our product (VisualWorks and ObjectStudio) nearly as soon as they are posted. I come across references to me just as fast. It's a new world out there, and there's really no such thing as security by obscurity anymore.

I've been meaning to get a screencast or two up, but a few things got in the way. First, I've been working with the folks at BlogJet and ecto to get the MetaWebLog API functioning for Silt, and that's taken a fair bit of my time. Meanwhile, two updates for BottomFeeder wouldn't load, so that entailed a new build (which I'm uploading now). While doing all that, I noticed that the way comments are handled on the server doesn't mesh well with the new comment tool in Bf. More mods. meanwhile, I've had a flood of email from various people asking me to blog a few things.

You get the idea - busy to overwhelmed (never mind my actual Product Management job :) ). Hopefully, things will settle down soon.

Chris Petrilli punctures a few lies that developers like to tell themselves

Here's the rules:

• Make it work
• Make it right
• Make it fast enough

That's it people. Premature optimization is the mental masturbation of geeks obsessed with complexity over anything else. If you think your site is going to have performance problems: You are wrong. You aren't writing Yahoo, or Google, and if you are, you can solve the problems incrementally as they happen, and as you understand them. Write in whatever language you feel most comfortable and productive in. I know quite a few Python (either Zope or SkunkWeb) based websites that are running millions of hits per day. That puts them in the top 1%. Your site isn't like that. If you think it is, you're wrong. Maybe someday it might be, but deal with it then.

That's about the size of it. I hear people talking about their language selection based on performance rationales all the time. Here's the thing - unless you are doing hard real time, that's just stupid. Read Chris' post. Repeat as often as necessary.

As you might probably know, we organize a squeaknic this Saturday in Bern. It is important that people who want to participate register themselves on http://kilana.unibe.ch/smalltalkcodingparty12-03-05/.

Up to now, there are only 6 people who want to attend. It is likely that we will postpone this event.

Here is the previous announcement:

Hello all,

SSUG is organizing a Smalltalk Party/SqueakNic/Coding Party. We invite all Smalltalkers to join this event to share their enthusiasm and knowledge about Smalltalk.

• WHAT -- SMALLTALK EVENT
• WHERE -- UNIVERSITY OF BERN - IAM Bern, Switzerland
• WHEN -- Saturday 12th of March 2005 -- 12pm until ...
• CONTACT -- email - bergel@iam.unibe.ch
• tel -- +41 31 631 3568 (fix phone), +41 76 58 56 323 (handphone)
• from Switzerland 076 58 56 323

Bring your lunch with you and contact me so that we can arrange something.

An apero sponsored by our sponsors will end the day.

• Daedalos Consulting AG
• iFace AG
• Object-Oriented Limited

If you plan to come, please register yourself on: http://kilana.unibe.ch/smalltalkcodingparty12-03-05/

Here's some information on how to get there:

• http://www.bau.unibe.ch/fakplaene/phil_nat/iam.html
• http://www.iam.unibe.ch/~scg/Resources/Maps/index.html

Gordon Mohr punctures some of the more starry eyed myths surrounding "Super Size Me":

I've got an idea for a film. It's like Super Size Me, except instead of eating all meals for a full month at McDonald's, I'll be eating all my meals at 5-star French restaurants.

I think the word here is "moderation".

This post from Ed Foster explains why I've always opposed schemes for product activation - they make an assumption that the end user is a thief, and force him to prove otherwise. That's not a good way to start a relationship with a customer:

"I purchased Adobe Acrobat Pro 7 last month," a reader recently wrote. "Several times since installing the software I am prompted to reactivate the product. After three successful Internet activations, I was directed to call Adobe. The person who answered the call accused me of installing the product on several PCs. I assured him that I had not done so. After reviewing my PC configuration he told me that activation does not work on RAID disk arrays. I had to install a non-RAID drive to allow Acrobat to activate properly."

Try as he might, the reader couldn't find anyone at Adobe who would offer a better solution. "I was forwarded to tech support who determined there was no workaround for this problem," the reader wrote. "Tech support's only suggestion was to purchase a volume license disk, since it does not have the activation 'feature.' They forwarded me to sales. The sales department would sell me a volume license CD, but I would have to pay a full volume license fee even though I only want one working copy. I asked for a supervisor and, after discussing the problem, he stated that I was not the first person to have this issue. He escalated the issue to 'upper management.' Two days later I was informed there was no solution for this issue. How frustrating."

That's a lost customer. The assumption made here is that activation prevents fraudulent use, thereby bringing in money. But does it? The customer in question had a normal setup (RAID is only going to get more common), and was given stupid responses (just pay us more money to make the problem go away). Look at what the end result of this is for Adobe:

• This customer was lost (read the whole article)
• They got bad publicity - Ed Foster, this post, additional word of mouth on it
• What did they gain?

When someone from management approaches you and asks for product activation, or cripples, or timebombs, ask yourself just how much negativity that will bring back - then try to argue against the scheme. It's not worth it.

After a lot of work and testing, I think I finally have the MetaWebLog API working on the production server. I've tested with both ecto and BlogJet - BlogJet now works against my server, fetching posts with categories. ecto fetches posts and categories, but doesn't match them up. I've been working with folks from ecto and BlogJet - they have both been very responsive and helpful. The problem is this sorry excuse for a spec from Dave Winer. He's periodically said he doesn't feel like he gets enough credit for this stuff; if I were responsible for that crappy API, I'd stay very, very quiet. Every publisher and blog server vendor has had to make guesses about how it works, because the "spec" is pathetic. In my server, I'm returning categories under three different names now - a guess based on the spec, one way for ecto, and another way for BlogJet. The only saving grace is that clients expect a dictionary, so they pretty much ignore data they aren't looking for.

None of this should be necessary...

So I'm merrily typing up another rant on the pitfalls and pratfalls of the MetaWebLog API when my network connection goes dark. The last time this sort of thing happened it meant a wonky router, so it was with a bit of trepidation that I went down to the basement - whew, the cable modem has the one light o non-connectivity going. Ok, I call Comcast. After a few minutes on hold, I get an agent - who tells me that there's a scheduled outage.

Hmm. So... let me get this straight. Comcast is my ISP. They have my email address - and they couldn't send me (and everyone else affected) an email? The website had an announcement, says the rep. Right, like I ever visit the Comcast website. Sure, that's a good place to put it - but additionally, how about a flipping email notice?

Is thought just completely dead in customer service departments?

.NET - not ready to replace a Smalltalk product that was moribund in 1997. Heh.

I think I've finally figured out the MetaWebLog and Blogger API (at least, for a couple of tools. Gosh knows what other clients expect). I now have both my own posting tool and ecto (for Windows) working against the MetaWebLog API. This is progress, and should allow Silt users to select whatever blog client tools they want. I figure I'll have more work against new clients as they come in, but this is progress.

Population of One links to an interesting story on the prospects (or lack thereof) for post graduate students - this certainly sounds like what my brother in law has gone through - years of being jerked around, with no real prospects of a tenure position. He's in bio-science, so the industry prospects for him are a lot better. Probably something that any person contemplating a life in the University ought to read as a cautionary tale, at least.

Well, the AutoLink debate continues to descend into unintended irony. Look at Steve Gillmor's rant on the subject - after this ode to letting end users do what they want with content:

Luckily, great chunks of the Sixties were captured on tape. The record companies are the Boomers’ WayBack Machine. Actually, it’s the record owners. Not just the official releases but the unofficial ones, the cut-outs, the outtakes, the remixes, the mono mix of Sgt. Pepper (reprise). The Dead’s thicket of tapers clustered around the soundboard. The WiFi junkies clustered around the power strips at the back of ETech two years ago. J.D. Lasica standing like a traffic cop with a radar gun at BloggerCon III.

He then comes out... against allowing end users to do what they want with content:

How do they cross the chasm Microsoft seems congenitally unable to ford? By cooperating on a standard link arbitration that lets users choose which variant or composite of the service they want. Join in a public-facing dialogue to establish an API for addressing link rendering, so that Google’s AutoLink can be chosen or spliced as a service with other offerings. Invite Microsoft to join in, given their SmartTag prior art. Then invite MyYahoo, who may be hard pressed to join in given their Roach Motel leanings around attention metadata. In ten minutes, your messaging goes from hailstorm to brainstorm

Which part of "The Google Toolbar is an optional (IE only, I might add) add on" doesn't he get? Which part of "AutoLink can be turned on or off" is hard for Steve? Finally, which part of "Microsoft will compete with Google on this" is confusing?

The cognitive dissonance is getting too loud for this hour of the morning...

Update: For over the top cluelessness, don't miss this stuff, which Dave Winer pointed to. The "panelists" are all crying for a way to opt out. Here's a big cluestick for these morons - you don't have to install the Toolbar. By gosh, you can uninstall it. Heck, you can use Firefox and completely ignore it. Oh, and here's another thing - I installed it this morning, and what they hey - when hovering over a Google created link, there's a tooltip that clearly identifies it as a Google link. It even changes the cursor, just to make sure I know that it's extra information.

The whole thing drips of the same arrogance I ranted about here - the professional media, academics - and now the "A-Listers" are all in a tizzy that the hoi polloi can - by gosh - remix the content and garner additional contextual information (no wait, that wouldn't be the semantic web, would it? Nah, it has to be complex and involve RDF for that). You know what? These people are (almost) torquing me off enough to make me want to use IE and AutoLink just to spite them.

Let me point out what I'm talking about: Here's a screen shot of an Auto-Linked page (after I pressed the button - by default, it doesn't mark up pages - you have to enable that or press the button). I'll use small words for the A-Listers - the user is in control, and has to ask for the enhanced content.

See that? See the clear marking of the link as a Google addition? You can't see the cursor change, because the print screen doesn't capture it. It's there, trust me. Now, let's have a look at the same page in Firefox:

There's the same page in Firefox, without the link. I'm simply horrified - I haven't been given any clearly marked additional content. Sheesh. Winer, Scoble, Doc (et.al.) - go play with Gorman. You can end up in the same place he wants to go, where experts only control the horizontal and the vertical, and the rest of us are banned from using tools we have to load on our own and set up. Can't have that - might empower us or something

And to add insult to injury, the "Bad News" site with the bozo video completely hosed down Firefox.

I've been working with James to try and get tools other than my own posting tool working with the Silt server. This has involved looking at the truly bizarro world of Weblog APIs. There are a few big problems:

• The specs (if you can *cough* call them *cough* specs) are very, very inexact
• Different posting tools (and servers, I guess) have interpreted these specs differently
• Some posting tools send any old api call up to the server, whether it's part of the API they claim to be using or not

For example - I grabbed a demo copy of ecto in order to do some testing. Just figuring out what it expected from metaWeblog.getCategories was an adventure. You go read what Dave Winer thinks is a spec, and see if you get it right. I Good luck - he makes some vague statement about it being a struct of structs; it should be an array of structs. What should be in said struct? Gosh knows, I had to dig through many websites in order to make intelligent guesses. Gads

I got through that mess, finally. Next, I was getting fault codes in ecto, even though posts were going to the server. A quick debugger session solved that - for completely inexplicable reasons, ecto sends mt.setPostCategories after a newPost. What the heck is up with that? That's Moveable Type, and that's not part of the MetaWebLog API. I implemented it anyway. Fine, that works. I still get some bizarre client message from ecto, a Dialog with the message: "Attempt to serialize data with a null reference". Not a clue what that is.

I know what insanity is now. It's the various extant WebLog APIs. Atom, you say? Sure, like those guys will ever get anywhere. They got bogged down in the non-problem of syndication formats, and are currently arguing over how many angels can dance on the head of a pin. Sigh...

There was an insidious little bug in the Silt blog server - updates to posts that changed categories didn't get visibly reflected in the sidebar - i.e., a category search would turn them up, but the count was always off. That was due to a caching error on my part - when updating the post, I was re-caching the old category instead of the new one. Dohh. This server is updated - if you are using Silt and want to update it, get the new code, load that, and then execute this snippet in a workspace (or in a headless server, file it in if you've left a path open for that):

| blogNames |
blogNames := Blog.BlogSaver default keys.
blogNames do: [:each | saver |
	saver := Blog.BlogSaver named: each.
	saver cache setupSearchCategoryCache].

That will get all the counts right, and the new code will keep them that way.

I've modified the comment tool in BottomFeeder to use the XHTML editor as well. Like the blog poster, it can be toggled back to plain mode. The Wiki style markup still works for both tools, if you have those settings on. I'm finding that to be a lot less useful for most of my posting work. Things are progressing nicely towards the 3.9 release with all this. The main tasks:

• Create the release notes that list the major changes
• Run through a few issues in the blog/comment posting tools that should be fixed
• Make the blog poster international aware
• Make a new candidate build

None of that should take that long - I expect to have a 3.9 build ready for release within a few weeks. There's some very good news for BottomFeeder coming from the WithStyle guys as well. One of the issues we've had with libtidy is stability - it's a C program, and some content can send it off to hell. Word is, the WS guys are creating a Smalltalk version - which will work on all platforms (something we don't have with libtidy), and will be stable. That will be really great news!

I see that the kerfuffle over AutoLink rages on - some people are now claiming that it's a possible legal issue. See this post, for instance:

From a legal standpoint, AutoLink looks questionable. The tool modifies publisher’s web pages by adding hypertext links without the publisher's consent. While this modification isn’t a huge change, I could still see some (many?) courts treating them as unauthorized derivative works. Honestly, it seems like a fairly routine copyright infringement. Google appears to be trying to position this as a situation where it’s merely acting as an agent for user instructions, but I’ve just recently blogged on how courts frequently slice through that argument pretty quickly.

Umm, let's see now. The marking up of the page happens on the client, not the server. Additionally, it's optional - very much so. First, you have to go out of your way to install the toolbar. Then, you have to enable AutoLink. So let's be very clear here - this almost certainly falls under fair use (at least, under any reasonable facsimile of it).

Based on the way courts act, there's no telling what they'll actually do, of course (I'm continually amazed at how the simple text of the 1st Amendment in the gets manhandled by US courts, for instance). In any event, I really don't understand the objections. This stuff is optional, under end user control, and it doesn't modify the server content. If this is a violation, then heck - so are highlighter pens and notes entered into a book.

I've been doing some more work on BottomFeeder this afternoon. If you've got the latest (7.3 based) build from the dev downloads, then grab all the available updates (you can have them load without restarting). Have a look at the comment tool - it uses the same XHTML editor that the blog poster does. If you prefer plain text editing, you can flip it over from the "View" menu - and you can set it to either way by default - look on the UI page of the settings tool.

I've pushed up an archive of the latest documentation (Thanks Rich!) for BottomFeeder - navigate to the download page, and grab the documentation archive. Then uncompress in the BottomFeeder main directory - replacing the current "UsersGuide" and "tutorial" directories.

A Slashdot reader asks:

"My friends and I have recently been in the market for a good new boardgame or other tabletop game. We have worked through the gamut of games like Axis & Allies, Supremacy, and War! Age of Imperialism. More recently we have been playing tile based games like Carcasonne and Settlers of Catan. I am looking for some suggestions on some new games we could get into."

Well. The game my group has latched onto is Puerto Rico. It's a great game, can be played in 45 minutes to an hour - and has better "staying power" than any board game I've ever played. Heck, if you do get to know it too well, there's an expansion set of buildings available. If you like board games and haven't tried it - run, don't walk to your nearest game site and pick it up.

Chris Petrilli hits on the main reason that so many people like the movie "Sideways":

Then it hit me. The reviewers see themselves in the character of Miles Raymond (played by Paul Giamatti), or at least I suspect. A slightly pompous middle-age exterior that masks a self-destructive interior, and a passion for a semi-obscure subject to which they dedicate a disproportionate number of words, and especially bizarre “edam cheese”-like phrases. (Yes, I know I succumb to this sometimes.) Otherwise, the movie was not “great,” simply very good. Too many issues are simply wiped away at the end and “resolved” without resolution. And more than anything, an important question is never answered… why would two people, the main characters, who were roommate in college, still be friends after so many years when they are so completely different.

Seeing aspects of yourself in the Miles character definitely explains a lot. I saw Sideways with my wife yesterday - and I have to agree with Chris - it's a good movie, but not a great one. Still, it made me reflect. I have to say that I think I've made something of a "mark" with BottomFeeder, so I'm not where Miles was towards the end (but not the very end) of the movie.

I found the movie both fascinating and painful. There were points in the movie where I wanted to duck out (Miles' drunken call to his ex-wife, for instance). I stayed interested though. I agree with Chris that there's one big hurdle to overcome with "Sideways" - the friendship between Miles and Jack. To relate it to things I know, Miles was the "developer", and Jack was the "sales rep". Believe me when I tell you that those two archtypes rarely go on vacations together.

There's another thing too - I don't know that many guys that have the kinds of heart to heart talks that Jack and Miles had a few times during the movie. Women have those conversations - most men just don't.

Having said all that, I enjoyed the movie. My take away thought - in very rough terms, there are a couple of forks in the road for guys when they reach the "middle years" (roughly speaking, the late thirties and forties). Some guys have a crisis where - like Miles - they despair at not having accomplished anything "big". They get divorced and/or buy a sports car, probably have an affair. Then there are the rest of us, who manage to be happy with what we've done - and don't end up like Miles (or worse).

Web Services are definitely too complex when even a Cxx type like Jonathan Schwartz's Weblog notices:

Web services may collapse under its own weight. No one at the conference said this. Those are my words. I'm beginning to feel that all the disparate web service specs and fragmented standards activities are way out of control. Want proof? Ask one of your IT folks to define web services. Ask two others. They won't match. We asked folks around the room - it was pretty grim. It's either got to be simplified, or radically rethought.

When you try to re-invent CORBA on port 80 over HTTP, bad things are bound to happen. Like, say, binary XML...

Spotted in MemoRanda

When Henry Ford famously adopted a 40-hour workweek in 1926, he was bitterly criticized by members of the National Association of Manufacturers. But his experiments, which he'd been conducting for at least 12 years, showed him clearly that cutting the workday from ten hours to eight hours - and the workweek from six days to five days - increased total worker output and reduced production cost. Ford spoke glowingly of the social benefits of a shorter workweek, couched firmly in terms of how increased time for consumption was good for everyone. But the core of his argument was that reduced shift length meant more output.

I have found many studies, conducted by businesses, universities, industry associations and the military, that support the basic notion that, for most people, eight hours a day, five days per week, is the best sustainable long-term balance point between output and exhaustion. Throughout the 30s, 40s, and 50s, these studies were apparently conducted by the hundreds; and by the 1960s, the benefits of the 40-hour week were accepted almost beyond question in corporate America. In 1962, the Chamber of Commerce even published a pamphlet extolling the productivity gains of reduced hours.

But, somehow, Silicon Valley didn't get the memo.

Now, I'm somewhat hypocritical bringing this up - anyone who hangs out on the Smalltalk IRC channel knows how many hours I tend to put in. There's a difference between that and crunch mode though - in my case, I'm working of my own free will - no one is forcing me to (heck, my management chain usually has no idea what I'm up to :) ). Crunch mode, as used by a lot of development shps, is not the same thing. With all the management and process fads out there, you would think that someone would try to correlate hours worked with bug rates...

Here's a fascinating article on the history of Eclipse at IBM. I got a laugh out of this explanation from Lee Nackman:

IBM turned the idea of a tools platform over to a subsidiary, Object Technology International, in Ottawa, which used small teams to develop new tools. IBM's own Visual Age toolset was based on the Smalltalk language and "was getting increasingly brittle." The new development environment would have to remain flexible and allow dissimilar tools to plug into it and share files.

Increasingly brittle? How so? I love the way he makes a sideways swipe at Smalltalk that way without really explaining what he meant. Oh wait - I think I get it - Smalltalk didn't require an army of consultants from IBM Global Services, or the large license fees they charge for WebSphere. That's what he meant by brittle - it worked for customers, but not for IBM sales reps.

So tell us Lee - what was brittle about your Smalltalk product? Inquiring minds want to know. If it is brittle, and you want to help your customers out - there are committed Smalltalk vendors around.

Hat tip to Jason Jones, who sent me this link

The development builds of BottomFeeder have had a problem getting updates - the downloads have been failing. I took a look at this just now, and realized that I was being hit by a change in the HTTP client libraries in VW 7.3. The files are being downloaded without any encoding information, so they default to iso8859. The trouble is, that's not what this data is - it's binary. For reasons I'm not entirely clear on, smaller parcels download fine - it's just bigger ones that have trouble. What I had to do is tell the client not to decode the content - which keeps it in binary form. Dump that to disk, and it all works. So - to get updates (if you have one of the new, 7.3 based development builds), do the following:

• Change your upgrade url so that the number is 73, not 721
• Check for updates and download PatchFileDelivery (this should come down)
• Quit BottomFeeder, restart
• Now you should be able to download any updates you want - and load them on the fly, as per usual

I'll be doing a new build this weekend to rid the downloads of this issue. I'll also be including the new docs, which I just got from Rich Demers. I have some more work to do on the post tool, and I intend to move the comment tool over to support WYSIWYG (like the post tool). Stay tuned.

I saw a question about XML editors over on the Lockergnome Forum. I think they need to check out Software With Style. I'm using it to create this post :)

I'm in the process of updating the development build of BottomFeeder now. The current development build (this is the full build from the dev links on the download page) had a bug in the update tool - a bug that prevents updates from properly downloading. I've fixed that, and I'll update this post when the upload is complete and in place. I've been working on getting the Blogger and MetaWebLog support working in the post tool working - I'm getting closer, but I need to set up a test blog to work against. Stay tuned on that.

Update: The new dev build is up

Cincom's President, Tom Nies, has been interviewed by Smart Business Magazine. Check it out here

I have a nagging doubt about the lovefest over in the Java tools universe. This comment from the eweek article explains why:

Meanwhile, although Borland was a founding member of Eclipse, the company never based its core Java tools framework around the Eclipse platform. Yet, sources said Borland has now set its sights on the overall application lifecycle and may be willing to offer concessions on the IDE side of things. For instance, Eclipse could become Borland's core IDE, or the company could deliver an enhanced version of JBuilder or a JBuilder replacement based on Eclipse, sources said.

Borland could be looking at Eclipse in the same way it views Microsoft Corp.'s Visual Studio .Net. "In the same way, rather than trying to compete with Visual Studio, they are going to build around it," Murphy said.

Here's the thing - that kind of unification means that Java developers will get exactly one baseline toolset. Sure, there are tons of plugins, and there will be more. The basic decisions have been made though, and that's it - no possibility of a different vision for development. Standards are nice, but they tend to freeze innovation - and I don't really think that Eclipse represents the end all, be all of software development. It's a distinct step back from what IBM had with VisualAge, for instance.

Eduardo Pelegri-LLopart thinks that Binary XML is inevitable:

One of the arguments against the need for a binary encoding of XML (like Fast Infoset) can be summed up as: "just wait until the technology catches up", or maybe "Moore's Law makes Binary XML unncessary".

Although that may be true in what some people could describe as "traditional" applications of XML, there are many legitimate use case of XML where this is not so. Many of these use cases have appeared as the market wants to take advantage of the benefits of XML in new fields. There are a number of reasons, some economical in nature, some technological, underlying these use cases.

The main argument against binary xml is that it's an oxymoron. Pretty much by definition, XML is a textual interchange format. You want binary? There are some fine choices around that already work - ever heard of CORBA? And before I hear about CORBA's supposed faults in contrast to *cough* WS* *cough*, I'll point out that the only reason XML got traction where CORBA didn't is port 80 and HTTP.

Yeah, we need another binary RPC format... like we need a hole in the head.

When this story applies to you, it's time to step away from the PC:

You're in the middle of a frenzied fragfest when it hits: You gotta pee--bad. Whatcha gonna do? Getting up from your computer clearly isn't an option--any 733t d00d knows the deathmatch owns the bladder.

Enter the Internet urinal, a handy-dandy portable pee device marketed specially for the PC-bound. Each contraption is made of hard plastic, comes with a "female adapter" and holds 32 ounces--a whole lotta recycled Red Bull.

"With the Internet Urinal, you'll never have to leave your computer again," touts a promo on ThinkGeek. "Imagine the freedom--destroy your opponents in that all-important 'Quake 3' clan match without taking a break; drink as many cans of BAWLS as you want and still be able to make that last important trade before the market closes."

Time to get a life :)

This morning, I had a rant about the state of blog posting APIs. It's worse than I thought then :)

Have a look at the MetaWebLog "spec" page, for instance - I have no idea what data a server is expected to return, nor do I have any idea what a client should expect to see. I can code a client defensively, but a server? I've been testing on the IRC again, and the tool that was hitting my server couldn't handle some of the data I was sending back. This is just maddening. I mean really - just marvel at this excuse for a spec:

In newPost and editPost, content is not a string, as it is in the Blogger API, it's a struct. The defined members of struct are the elements of <item> in RSS 2.0, providing a rich variety of item-level metadata, with well-understood applications.

The three basic elements are title, link and description. For blogging tools that don't support titles and links, the description element holds what the Blogger API refers to as "content."

For tools that don't support titles and links? I'm supposed to know which ones those are.... how? Can't we find a happy medium between the undefined crap that we have now and the over-defined insanity that is Atom?

The Wall Street Reporter interviewed Cincom's President, Tom Nies recently. Here's the interview:

WSR: Maybe we could start off with a brief history and a general overview of the company.

NIES: Cincom’s been operational for 36 years. We operate all around the world and have three to four thousand strategic customers. We deal with commercial, Government and institutional buyers. We don’t sell consumer software. We compete in the marketplace with firms like Oracle, SAP, and Siebel, selling not only strategic software applications and software products, but also application development, database management and software and services that clients use to build and develop their own applications.

WSR: Tell us about some of the emerging and developing trends that you see within the industry and explain to us how the company’s products are redesigned to capitalize and address these trends.

NIES: I think the most important trend is that customers are becoming much more demanding buyers. There is a significant supply of excess software in the marketplace today. I estimate as much as 40-50%. For any type of product or solution one wants there are half a dozen good potential providers. This gives customers the opportunity to demand more for their money. As a result, software implementations are now a better buy for them than they were in the past and will be even better in the future. Software companies who provide their customers more value at a lower cost, with more rapid ROI, while minimizing risk, are going to benefit handsomely in the future. Those who continue to require very drawn out and significantly excessive costs of implementation and support will suffer. Simple enough. In a buyer’s market, customers will demand and get greater value at much lower overall cost. Only naïve buyers will accommodate vendors in today’s marketplace as they did in the years leading up to 2000.

WSR: One of the developments we have seen in recent times is the whole regulatory requirement of Sarbanes-Oxley. And companies are looking to such areas as business intelligence software. How do you see this particular trend developing and how is the company addressing this?

NIES: Sarbanes-Oxley is a good indication of the fact that investors want more information about the business just as management needs more information about the business. There has to be a lot more integrity in the figures and facts because excessive risks will no longer be tolerated. But besides knowing more about the business and reporting to the owners and the regulatory bodies, the globalized world we are living in today has increased the competition so much that more effective, better and more comprehensive use of software in just about every area of the business is absolutely mandated. That is one of the reasons why I think we are going to see a great blossoming in software opportunities in the future, certainly in business intelligence. It’s not one that Cincom is now promoting heavily, but it’s a new opportunity area for us too.

WSR: In terms of partnerships and alliances within the industry, how does Cincom use them to further your objectives?

NIES: Partnerships are key. One has to minimize the cost of selling and distributing software, as well as broadening the base availability and distribution of software and solutions offered. So, we not only look to allies to supplement and round out our product line, but also we look to partners who would use our software line and some of the services we offer to better satisfy their customers. We are working heavily with developing and expanding partnerships. We originally built the company around a partner-related environment, and I think this is a way forward for most companies today. It’s a major trend in our industry to see expansion of partnerships everywhere.

WSR: Cincom competes in the industry against such leading companies as SAP, Oracle, and Siebel Systems. How does Cincom distinguish its technology and its product offerings from these competitors?

NIES: Today, almost all the software providers offer the customer more than they really need to satisfy their requirements. So, emphasis on product feature and functionality is an area of marginal utility and with diminishing diminishing returns to boot. To win more consistently in the marketplace today one must deliver more value to the customer at a lowered cost. The differentiator we show companies from an Oracle or a SAP is that Cincom will implement a system of similar capability for perhaps a fourth or fifth, maybe in some cases, a tenth of the total costs and half to a third of the time required to reach operational utility of the systems desired. That’s our value proposition. Cincom does this consistently and we believe that is the significance of our comparative. Advanced in more value, delivered for less time, cost and risk are having an impact for Cincom in the marketplace.

WSR: Can you tell us about the background and experience of the management team on board?

NIES: One of the great strengths of Cincom is that our company has proven to be a very attractive company to our associates. We not only are able to attract good managers and top executives to Cincom, but we were able to retain them. We have an average of 12 to 15 years or so leading and guiding every one of our business pursuits. Our people are committed to our business, and Cincom is committed to our people. We have probably the longest average employment rate in the industry. We also have the highest returnee rate. One out of every 12 of our people are people are returnees. Over 25% of Cincom’s staff have more than 15 years with our company. We are deep in talent. We have a really experienced and zealous management team. We can, and do consistently, deliver on our promises because of the skill, quality and experience of our people.

WSR: Cincom serves thousands of clients in six continents. How do you foresee the next two to three years— now—as a time of expansion and development for the company?

NIES: Customers want providers to take more and more responsibility. So, we are expanding our offerings. Hosting services, outsourcing services and more comprehensive facilities management types are now being provided to customers. We see this as a great growth opportunity for us because customers who are looking at the IBMs, Computer Sciences, EDS, and others are now looking for alternative suppliers who will provide the same type of quality and comprehensiveness of service, but at significantly lower prices. So, this is another market opportunity that we are pursuing with exactly the same model as we employed for our software product offerings. What Dell has done for PCs—that is to provide a very similar quality at a much lower price—is what we are doing in the area of software and outsourcing services. It is a model that we think will play well into the future; more value with lower cost. We see this not just with Dell, but we see these demands being made in almost every industry, in every part of our now globalized commercial world.

WSR: In terms of geographic expansion, what are some of your other new key markets that you think might represent an opportunity?

NIES: Asia-Pacific is absolutely on the top of the list by far. Europe and America are mature markets. They are good markets with moderate growth. And we are penetrating and developing these markets quite well. But, the growth rate in these markets is much less than the growth rate in places like China and throughout the rest of Asia. Japan is still a very good growth market, but China, India and much of Asia are now very, very substantial growth opportunities. That’s why just about everybody is going there as fast and as significantly as they can.

WSR: Perhaps you could give us just an idea of some of the key milestones that we can expect to see from the company over the next 12 to 18 months.

NIES: Throughout this entire decade, we have averaged over 80% compounded return on invested capital; 80%—that’s three to four times or five times what is typically generated among good performing companies in our industry. So, a very high return on investment and also significant increases in earnings per share is a key Cincom emphasis. We have increased earnings per share by seven-fold over the last five years and we are averaging, as I said, 80% return on capital investment. We are looking to expand the business significantly without any kind of adverse effect on these operating results. This will be no easy task for us. But, we are committed to high returns on investments for ourselves just as we look to provide our customers high returns on their investments with us.

You can view the entire report on the Wall Street Reporter site.

This is stupid, but here it is - if you are using the latest development version of BottomFeeder (i.e., you downloaded a dev build in the last few days), there's a bug in the update sub-system. The upgrade url in settings has to be modified as follows:

• Replace the text '721' with '73'
• Make sure that there's a '/' at the very end of the url

I'll be fixing those before I go to release, but you can get from here to there with those work-arounds.

I like this rant in the BileBlog - no one does rants as well as Hani. Here's some of the milder stuff I'm willing to paste here - but go read it for a few devastating (and hilarious) take-downs:

I think one of the flaws of Mark's talk is that he's forgetting (or is unaware of) his audience. They aren't, as Floyd would like to think, clever leader types. They're just everyday grunts who have enough spare time and meaningless enough jobs that they can fart off on TSS every other day, interspersed with the odd person who has been sufficiently beaten with the cluebat.

The whole SOA myth makes for a great sales pitch by IBM types to high level 'architect' types whose job involves little more than doodling with crayons and going on IBM sponsored golfing trips. It does not, sadly, translate well to gruntspeak. us grunts are simple folk, we like code examples, we like concrete classes, and by god, we like xml. Anything else and most of us will be flailing about helplessly trying, and failing, to relate to the subject matter.

Read the whole thing, as they say

I've had harsh words to say about Atom in the past, but that was mostly over the feed format. I haven't looked at the posting API yet - maybe I should. The Blogger API and the MetaWebLog API are simply nightmares. There doesn't seem to be any standard way for client tools to interact with a server - I was debugging the interaction between a client and my server last night via IRC. Even better - the client was set to use the MetaWebLog api, but was sending requests to blogger.apiNameHere names. Sheesh. There was also an interesting difference in api points - I had implemented 'getUserBlogs', and the client was sending 'getUsersBlogs'. A quick Google search turned up references to both. Sigh.

I implemented both names, pointing to the same method. I had to map blogger names over to MetaWebLog entry points, at least for the tool being tested last night - who knows what oddness will turn up next. What a complete mess...

Well, I knew that this was coming. I just didn't realize that it would be coming from something that pretends to be a news source:

Photo editors cropped her head onto a model's slimmer body to create the visual effect, which even the New York Post knows is an ethical black hole (err, maybe they don't). A footnote does appear on page three with the credits: "Cover: Photo illustration by Michael Elins ... head shot by Marc Bryan-Brown."

But that's not exactly Clarissa Explains It All for the average reader. Another Jennifer Aniston on Redbook, you say? So do we, even if assistant managing editor Lynn Staley believes "Anybody who knows the (Stewart) story and is familiar with Martha's current situation would know this particular picture" was a "photo illustration."

Yes, these fakes tend to get picked up quickly by attentive readers. However, how many casual readers hear about that? And yes, this particular case is trivial. I'm just waiting for the first political dirty trick launched using photo/video editing - it's a matter of when, not if. The bottom line - you simply can't trust photos, video, or audio anymore unless you trust the source. The thing is, news sources are tossing their believability down the tubes with stunts like this.

Ten is a good number notes a problem with mathematical education in the US:

In 2000, the state with students with the best mathematics proficiency percentage was Minnesota with 40%. That means that the best we could do in 2000 was 60% of 8-graders unable to apply mathematics to real-life problems. This is a sorry state of affairs.

He goes on to list many disturbing statistics that show just how innumerate most people end up. Towards the end, he adds a link to the sorry state of textbook production, implying that this is the biggest problem.

It might be the biggest problem. However, it's not the main reason (IMHO) that students end up having no practical mathematical skills. Let me run through the laundry list that I have:

• Calculators introduced in third grade
• No emphasis at all on basic computation skills
• An over-reliance on amorphous "computer skills"

I was very upset to find the local schools having the kids use calculators as early as third grade. Most students hadn't memorized basic multiplication (or even addition) facts; the school system seemed to think that "dull", and just handed out calculators. My wife and I had to do the drill work ourselves. Now sure, in "real life" you'll always have access to a calculator. But if you can't do basic computation, a lot of high school and college level math is really tedious. Go out and test anyone who's in their 20's (or younger) to get a feel for just how bad it is - now consider how they are going to make sense of whether a given sales price is of any value. If they can't do that, then they certainly can't make sense of political debate centered around budget figures.

What we've got is a completely innumerate voting population - which is every bit as dangerous as an illiterate one. It's not taken seriously though - do you ever see anyone making light of not being able to read in a movie or TV show? How many characters do you see saying "I'm no good at math" - or, on the other hand, why is it that most of the mathematically literate characters are portrayed as complete losers?

So yes, the way textbooks are prepared is a problem. There are simpler problems though, and yes: I'm willing to lay this one directly at the feet of the schools and the teachers. They know full well what they aren't teaching in this area, and there's no good reason for it.

I've been making major changes to the posting tool (full support for the Blogger and MetaWebLog APIs), and I've fixed bugs in the interface between Bf and the posting tool. As well, the current dev build can't actually download updates. Argh! I'l have a new dev build up tonight; anyone using the dev build should grab it (i'll update this post when it's ready)

Update: The new dev build is ready for download. Scroll down to the dev links

I understand the porn and poker spammers - they aren't "respectable" businesses, so they don't really worry about their reputations. It's different for a vendor like these guys: www.thebiggestdeal.com. It looks like someone ran a bot on a server owned by a business partner, sending referer spam out. It's unclear whether that partner did it, or had a machine "owned" - but it looks like my initial assumptions were wrong. I've been exchanging email with Bill White, the President of the company - he sounds like a standup guy. The spammers do damage wherever they go

I think I've got the Blogger API and MetaWebLog API sorted, both for the client posting tool and for the Silt server. I'll be pushing updates after lunch - I need to take my daughter to the Orthodontist now.

It seems that when I put together the blogger and metaweblog api support for the post tool (2 years ago, I think), I didn't really read the specs (such as they are) correctly. Mind, you, these api's are a rats nest of oddness - semi-standard entry points like getUserBlogs, for instance, seems to be required by various tools, but is only anecdotally documented. In any case, I'm in the process of going through the weeds of these apis. I should have something semi-reasonable later today

It's official - Smalltalk is less evil than Java :)

Smalltalk:

Java:

Meanwhile, it seems that Smalltalk advocacy is a good thing:

I've updated the ssp pages and style sheets that Vassili sent me, and gotten everything in place. The changes are visible on the main page, the archive page, and on the comment entry page. I haven't moved over the other sites, but that's just a small bit of configuration. I wanted to get mine up with the new stuff first, and make sure that everything looked ok before I went through another round of symlink fun

I found an interesting request in Scoble's blog:

But Yahoo's API doesn't look like they really gave me what I wanted.

Here's the first thing I wanna try to build: a search engine without blogs.

Huh?

Seriously. Blogs are increasing noise to lots of searches. We already have good engines that let you search blogs (Feedster, Pubsub, Newsgator, Technorati, and Bloglines all are letting you search blogs). What about an engine that lets you search everything BUT blogs? Where's that?

Well, explain something to me - how does a search engine differentiate a blog from an arbitrary website? It's not as if their labelled in some universal way (nor will they ever be). He then goes on to state that Yahoo's API "isn't good enough" to support that. Earth to Scoble: that might be because you asked for an impossible feature. Sure, we could have an engine omit things listed in those indices. The trouble is, it's not as if all blogs are listed in those indices. Second, there are things listed in those indices that aren't blogs - Feedster, for instance, indexes RSS and Atom feeds. I know that I've submitted non-blog RSS feeds to feedster.

I might as well as for a DWIMD (do what I meant, damnit) compiler...

There's been a fair bit of commentary from this post on Google's AutoLink. Here's the thing - people complain that Google's service is "evil" because content producers have no opt-out option.

Well. I hope none of the people who make that complaint ever use any of the following pieces of equipment then:

• VCR's that have commercial skip capability, or fast forward
• TiVO or ReplayTV (or any PVR), using either 30 second skip or ad skip
• Any music copying capability that moves songs from "their intended place on an album" to a tape, iPod, custom CD (you get the idea)

With all of those tools, the original content producer has no opt out capability. In fact, most of us - including many of the same people complaining about AutoLink - have raised a hue and cry (properly, IMHO) over the RIAA/MPAA attempts to kill off those capabilities. But hey - if you oppose AutoLink on the grounds that content producers have no ability to opt out, then you better be willing to bend over and take it from the RIAA and MPAA. Because it's the same issue. The only difference is the size of the entity protesting.

Update: Looks like Scoble better give up his TiVo. After all, it's just horrible that it provides a butler service, allowing him to view content in ways that the producers don't control. Ditto any MP3 players you have lying around too, Scoble. Heck - why don't we forbid anything but read only CD's and DVD's - that'll keep us safe from anyone who wants to shaft those nice content producers. I'm sure that they have our best interests at heart, after all. You can send me the TiVo Scoble - clearly, it's evil technology...

Well, this is why I call them dev builds. The update tool in the 7.3 based BottomFeeder build is broken, so I'm in the process of putting together a new build. It was a stupid problem having to do with changes to the Http client code that the upgrade package wasn't accounting for. I'll have a new build up later today

Update: The new build is up

Interested in the Smalltalk Solutions Coding Contest? Then register here - registration runs from now until May 13. We'll be announcing the contest itself shortly after registration closes.

Sometimes I'm just glad that I don't eat in truly high end restaurants. Scroll down on that link to this:

I just don't know what to say to you. A well done steak, particularly a filet, is a crime against god. Anthony Bourdain, who I consider to be a personal hero, said in Kitchen Confidential:

[steaks], if ordered well done were routinely thrown into the deep-fryer until crispy, then tossed into an oven to incinerate further ...

I cannot imagine how offended your waiter and chef were by being asked to destroy a piece of beef like that. Seared, with a cool red center, if you please. Well done? For f**** sake. I bet you like Pilsners and drink Corona with a lime.

If the chef and/or waiter were truly offended, someone needs to remind them of a simple truth - the diner is paying their salary. It's really not their problem how the diner orders food, so long as he pays. We have this same problem in the technology sector. We like our little holy wars over languages, operating systems, and hardware - and certainly we are entitled to make our pitches. The end customer is the one paying though, and we need to remember that.