Smalltalk Tidbits, Industry Rants

Interested in concurrency in Smalltalk? Then go and look here. As it happens, Adventa is a customer of ours.

If you have an older download - or CD - of Cincom Smalltalk non-commercial, then you'll get prompted for a CD Key when you run the installer. Starting with the last release, we removed the CD key. We were getting lots of problems with that approach, so we got rid of it. If you have an older CD or file set you need a key for, send me an email

Jonathan Schwartz thinks that PC's will run south in price and end up in the same boat as mobile phones - mostly nominal cost followed by long term subscription (broadband) plans:

And as I was saying on Steve's Gillmor Gang, give it a year, I'll make a big bet, you're going to see the same thing begin to happen in the PC industry. Will the average selling price (ASP) of a PC continue to meander south? Yes. Unrelated to component cost. Will it go to free, like handsets? Absolutely. In exchange, consumers will sign up for network plans, DSL, cable, you name it. On PC's, or more likely equivalent dedicated devices branded by carriers - just as in the handset industry, where carriers are increasingly deploying 'branded' handsets. It's been a tad tougher on PC's when they didn't control the software load, but it's just a matter of time.

There are a few problems with that - the biggest being inertia. Mind you, I think people would generally like the idea of being able to easily upgrade their PC's every 18-24 months (as they can do with cell phones) - on the other hand, I see no real sign of the big providers - cable or dsl providers - taking steps into those waters. Mainly though, Schwartz is either misreading the IBM sale, or trying to make it fit his preferred world view. The IBM sale is about something very, very simple - access to the exploding Chinese market for PC's. Consider:

Under the deal, IBM will take an 18.9 percent stake in Lenovo. Lenovo will pay $1.25 billion for the IBM PC unit and assume debt, which will bring the total cost to $1.75 billion.

The combined venture will have roughly 10,000 IBM employees and 9,200 Lenovo employees. It will be headquartered in New York, with operations in Beijing and in Raleigh, N.C.

IBM gets a major stake in a firm that has access to the Chinese market, and has managed to move their management team to the US. That will make it easier to work with them, while (they hope) maintaining access to the lucrative Chinese market. That market's appetite for PC's is growing much faster than the US market is - if this all works, IBM will rake in a nice sum of cash from Asian sales while not having to do a whole lot of work for it.

Now, there are obstacles to this - there are numerous stories out like this one, hinting at security related concerns on the part of the US government. It's not guaranteed to happen - but if it does, IBM gets a huge jumpstart over their competition in terms of access to the Chinese market. Meanwhile, Schwartz is - again - either misreading or attempting to skew the story to fit his preferred worldview:

To me, the sale of IBM's PC business proves the point - the company that invented the category (me, I like to believe they cloned Apple's business), is now exiting due to poor financials. And attempting to sell it to a company whose principle competitive differentiator is cost (and geographic/political proximity to a growth market). It's my view that operating a PC only business doesn't really make a lot of sense - at least in the long run (which is why Dell's moving ever closer to selling microwave ovens, and still trying to convince the world they can run their supply chain on 1-way Xeons - hm, note to self, someone should ask their CFO on their next earnings call what systems run their supply chain).

Yes, he's correct about handset sales outstripping the sales of PC's, and he's right that the margin there will only grow. I'm not convinced that he's right about who will benefit on the back end of that - I know of plenty of large outfits in the network space that are retiring Sun boxes and bringing in commodity Linux as a replacement. This strategy from Sun looks a lot like the one they ran (successfully) during the dot com boom - back then, the expanding number of websites drove a lot of hardware sales to Sun. I wasn't convinced then that their huge investment in Java was helping them a lot - those websites were going to go up regardless, and Sun could have sold a lot of hardware without the huge software investment in Java. Likewise, I'm not at all convinced that the prevalence of Java phones helps them a lot here, either. Even if they do succeed in selling lots of hardware, the simple fact is that said hardware would sell with or without Sun's Java investment - Java's existence is not a pre-requisite for the mobile phone space. The point is - the JavaSoft group is a huge brick weighing down Sun's business, because the hardware sales they hope to make are almost completely independent of its existence. At the same time, other vendors - IBM, for instance - have used Java quite successfully as a shovel to pile money into their maw.

Sun is as obsessed with IBM now as they were with Microsoft over the last decade. I suspect that said obsession will be every bit as useful for them.

There's been a lot of work on the XML to Object front in the latest versions of VisualWorks - there's a nice set of wizards that make it a lot easier to define and deploy a web services application. Our lead WS* developer, Tamara Kogan, asked me to post a tutorial presentation on this stuff - here it is. A small note - you will probably get some kind of warning about using IE when you browse that - I tested it with Firefox, and it works fine. I'd push up something in PDF, but I have no tools for that.

Update: As pointed out in the comments, many of the slides have issues in FireFox (and other browsers other than IE I'm sure, not to mention other platforms). I saved the slides as html from PowerPoint, and it looks like ActiveX is required. I'll happily make these available in a more easily accessible format if someone can tell me how to do that from the PowerPoint slides I have.

Update 2: See this post for the pdf versions I posted

Never mind the CAN-SPAM act - this NY Times article (registration required) explains the economics of spam. Which also explains why you won't see an end to it anytime soon - either in email, blog comments, or referer lists:

A spammer can often expect to receive anywhere from a 25 percent to a 50 percent commission on any sales of a product that result from a spam campaign, according to a calculus developed by Richi Jennings, an Internet security analyst with Ferris Research, a technology industry consulting firm.

Even if only 2,000 of 200 million recipients of a spam campaign - a single day's response rate for some spammers - actually go to a merchant's Web site to purchase a $50 bottle of an herbal supplement, a spammer working at a 25 percent commission will take in $25,000. If a spammer makes use of anonymous virus-enslaved computers to spread the campaign, expenses like bandwidth payments to Internet service providers are low - as is the likelihood of anyone's tracking down who pushed the "send" button.

With that kind of money to be made, it's no wonder the problem "sticks"...

Avi makes an excellent point about Smalltalk in relation to some of the other popular dynamic languages:

I don't know Python's internals anywhere near as well as I know Ruby's, but I suspect that at least some of the above applies to Python as well; certainly, there's a lot of C code in that camp as well. In Smalltalk, however, the only thing that's in C is a tiny virtual machine that understands a very simple set of bytecodes - it knows about pushing and popping objects from the stack, accessing indexed slots, blocks, and method dispatch. What it doesn't know anything about is the language syntax, standard library, compiler, debugger, profiler, thread scheduler, exception system, and so on, and so on - all of these things are implemented in Smalltalk. You want to change the way Array works everywhere in the system, well, find the appropriate Smalltalk code and change it. You want to extend the syntax of the language, subclass Compiler or write your own - the bytecode spec is easy to target. Have an idea for a new way to debug? There's plenty of code you can look at in the current Debugger, Process, and Stack classes to help you along, and when you're done just modify Exception to bring up your debugger instead of the standard one. The language doesn't support continuations and you want to add them? No problem; 10 lines of code and you're there.

That's the key benefit of Smalltalk - and it's a benefit over Java, C++ and C# as well - the fact is, the tools for those languages and systems are much harder to write than their Smalltalk analogs - and a large part of the reason why is that much of the system in question is hidden away from the developer. In most Smalltalk systems - certainly this is true of VisualWorks and Squeak - most of the system is in Smalltalk, not in the VM. That makes it a whole lot simpler to get in and muck around. Add in the productivity benefits of Smalltalk over the more common systems, and you've really got something.

But again, I'll say don't take my word for it - go grab Cincom Smalltalk or Squeak and see for yourself.

This strikes me as a "just desserts" sort of thing for the folks at Gartner - PR Opinions reports on a warning issued against re-upping Meta Group subscriptions in light of the merger:

(from Knowledge Capital Group) "We reiterate our recommendation to not enter into any new, nor renew any old Meta subscription based services until both Meta and Gartner management can tell you why you should.  Given what has been communicated so far, we feel that the chances of you getting what you want from the transaction and not having at least some of the value of the subscription vanish are very slim.  I know, it sounds harsh, but here is why we think this is just smart business..."

Oh, the irony...

I like the idea behind this post from Scoble:

But I'm watching the bloggers over at Sun Microsystems, Yahoo, Google, and here at Microsoft and I'm realizing that we have a lot of power. Howso? To do small things. So, I'm wondering, what are some small things that we could get done? Can we make a list of 100 small things? You know, things that a developer might be able to do in one day. Or, at least, less than a week. And that can be approved by someone seven levels down from the CEO. Like where I'm at. I probably can spend $500 without attracting too many MBA types trying to figure out why I'm becoming a cost center.

So I'll ask the same question as Scoble - what small annoyances in Cincom Smalltalk bother you? Send me suggestions, either in comments to this item, or in email. I can't absolutely promise a fix - as Product Manager, I'm trying to balance a lot of competing interests - but I can guarantee a flashlight on your issue. Try to remember - like Scoble, I'm looking for "small things"

Ed Foster explains why the "what me worry" approach to EULAs just won't cut it:

Unfortunately, though, there's something else we have to take into account here: DRM. Whether you think the acronym stands for Digital Rights or Digital Restrictions Management, it's an everyday part of life now. What hoops will you need to jump through to reload your software applications if you have to replace your hard drive? Can you cut that tune you paid 99 cents for to another CD without the RIAA storm troopers breaking down your door? And, given the leaning of industry and government, it's likely to be all the more commonplace. Will the broadcast flag on the TV show you're watching permit you to fast-forward through the commercials? Or will the RFID chip in your car keys lock you out if you miss a monthly payment?

So what's the connection between DRM and sneakwrap terms? The common thread is that with both, customers don't really know what they're buying when they put their money down, or at least not without reading through multiple layers of fine print. And they might as well not bother to do that, because the deal can be changed retroactively by the seller as you go along. Hey, you have been granted certain rights, up until the time we decide to take them away from you, at which point you have no rights.

For the vendors, of course, DRM also means never having to go before to judge to argue that their unconscionable EULA terms should be enforced, because the DRM is judge, jury and courtroom, all rolled into one. I don't think it's just a coincidence that the terms readers found most objectionable in last week's poll were from an adware company that is suing another adware company over its software being automatically removed by the competitor's software. The case is really one of EULA vs. EULA, with both companies claiming virtual ownership of your computer because you supposedly agreed to terms allowing them to do so. And, as spyware expert Ben Edelman reported from his research, it's quite possible for both programs to load through security holes without users having the chance to know it's happening, much less "click OK" to the terms signing over control of their computer.

This stuff is more than most people want to pay attention to - which is why the RIAA feels so free to just toss lawsuits around. I really, really don't like where this is headed, and I give a lot of credit to Ed for continuing to stay focused on these issues.

Rok Hrastnik has updated his RSS e-book - there are two new sections:

1. Using RSS for an increase in clicks, more linking, and better search engine visibility: Exact instructions on how to optimize your RSS feed and blog for an increase in clicks, linking and better search engine visibility. By Tinu Abayomi-Paul.
2. An interview with Ovi Crisan of 2RSS.com. Discussion on the distinctions between RSS and e-mail and how to keep using them together.

Rok is promising more new stuff to come - this is a good resource for marketing folks who have heard of RSS and syndication technology, but aren't really sure why they should be concerned.

Dare Obasanjo explains why Technorati's new tags are neither a new idea, nor anything particularly useful.

This is funny, but make sure to read the comments - the long one is just too much :)

Everyone knows that keeping customers happy is good for business - or at least, everyone pays lip service to the idea. Here's a post from Dale Wolf that goes into that using the phrase "customer love":

If we really loved our customers, cared about their issues and pains, their desires and wants, the expectations that they hold out for us 26 would we act a whole lot different that we do today? Could caring love change the corporate scene and produce more valuable marketing communications?

If we loved your customers as much as we love our companies, how would the content on our websites change? How would our service offerings get better? Is it possible that customers would love us back?

I think this starts with customer service. One of the easiest ways to torque an existing customer off is to offer sub-par service. The thing is, it goes beyond that one person - they'll talk. Take a look at this post, for instance - bad service spreads by word of mouth, and that word of mouth is accelerated by the blogosphere. If it ever was a no cost thing to tick off a customer, it stopped being one a little while ago - and ignoring the problem can have even higher costs. The very last thing you want is for a product search to list a bunch of complaints right at the top.

So what can you do? Take complaints seriously. You can't satisfy every customer or prospect, but you can treat them with respect, and you can chuck your script for dealing with calls. Nothing irritates me more in a service call than for the second and third person I talk to to be reading from the same script as the first. It's even more maddening when I get to a person who can work beyond the script and be reasonable, but only after investing an entire day of my life to get up to them.

People tend to be willing to go back to a place where they got good service - that's why we have favorite restaurants - and it's also why we have restaurants that we avoid like the plague. Truly awful food is rare (at least at US restaurants) - but you can find truly awful service. The same rules apply to business relationships - if you extend shoddy service, don't expect to see a customer come back, and don't be surprised when you learn that they migrated to a competitor. You don't have to "love" your customer - but you do have to be respectful and professional.

I have to hand it to Sony - their PR people are right out front, placing the blame where it belongs:

Kutaragi acknowledged that the button is less responsive than the others, in part because it's so close to the PSP's 480x272 screen. Because there isn't enough room to put the square button's detection switch directly underneath, it's off to the right, making it less responsive--and sometimes causing it to stick.

Nikkei Business reported that, to date, .6 percent of the 800,000 shipped units have been returned to Sony for repair. Kutaragi was unapologetic about the issue: "This is the design that we came up with. There may be people that complain about its usability, but that's something which users and game software developers will have to adapt to. I didn't want the PSP's LCD screen to become any smaller than this, nor did I want its machine body to become any larger.

"The button's location is architectured on purpose," Kutaragi added. "It's according to specifications. This is something that we've created, and this is our specification. There was a clear purpose to it, and it wasn't a mistake."

Somewhere in Sony's marketing department, entire rooms of people are grinding their teeth down to stubs...

Every so often when I get directions from MapQuest I just blink - the part of the route I understand doesn't always correspond to the directions they give me, and sometimes, they are just flat wrong. Last summer I tried to get directions to a girl scout site for my daughter, and MapQuest sent me to the wrong county. Well, it's not just MapQuest - looks like Microsoft's map tools suffer from some of the same issues. Get a load of this map:

"The suggested route starts in the middle of the Baltic Sea on a ferry from Stockholm to Helsinki. Then you drive around in the Southern part of Finland before, magically, driving across the Baltic Sea on 'local roads' to arrive in Tallinn," writes Toben.

"The distance is given as 68.4 miles (which sounds about right) and the estimated time for the journey is 4 hours and 36 minutes. The return journey is by the same route in reverse, only now it takes 4 hours and 53 minutes."

So we know the flying car will do scheduled pickups - so long as you're prepared to bob around in the North Sea.

The Register has a lot of fun with MS on this one, but my experiences with MapQuest tell me that it's not about MS - it's about crappy data coming in from somewhere. I rather suspect bad data entry with the all too obvious results.

Like Ted, I recommend the book The Pentagon's New Map by Thomas Barnett. It's an engaging book - Barnett lays out his worldview, and does a good job of it. I can't say that I agree with everything in the book, but I can say that it challenges assumptions - regardless of where you fall on the ideological spectrum. Read the book, make up your own mind. One thing I didn't realize is that Barnett has a blog.

Wow, it's certainly the end of an era. SBC is buying ATT:

SBC Communications on Monday announced plans to acquire AT&T in a $16 billion deal, a move designed to bolster SBC's sales to enterprise customers nationwide and give it greater access to national and global networks.

I found it ironic a few years ago when Mobil and Exxon combined (mostly undoing the Standard Oil breakup). It looks like history is repeating itself, only at a highly accelerated rate.

Ahh yes, it's that time of year again when the pundits come out and show off what they don't know. I came across this article in Salon on the Mac (hat tip - Scoble), and wasn't disappointed. Take this:

But that halo didn't seem to work. For some reason, in 2004, vast numbers of Windows people didn't look at their iPods and decide to buy Macs. Why not? Perhaps the answer lies in what Jason Snell, the editor of Macworld magazine, says is the essential difference between Windows people and Mac people: Mac people love their computers on a personal, emotional level. Windows people, on the other hand, prefer to think of their machines as office tools, gadgets no more special than the stapler. Windows users don't expect much in the way of quality, beauty or elegance from their machines; if they did, they'd be Mac people. Instead, they expect their PCs to perform a great many tasks, and they've resigned themselves to having to labor over those tasks.

This is the sort of thing that I really, really hope Apple's sales/marketing people don't buy into. If there's a Mac/PC person divide, it's a small one, and it doesn't explain the failure to buy. What does is things like my recent trip to CompUSA. As I expected, I got a stream of comments after that post, including a couple citing the elegance of the Mac. The problem is, that elegance buys me fewer features for more money. Wander down to a Target or Wal-Mart if you want to see how average people buy - they compare prices and features. Like me at CompUSA, they note the amount of memory, the amount of disk, and the relative price points. They then end up buying the PC. Clearly, this Salon guy doesn't get that - witness this:

Well, for one thing, the Mini's cheaper than the iMac was. It is still possible to buy a Windows machine that costs less than the Mac Mini, but you'd really be scraping the bottom of the barrel, and even if you got something with comparable computing power -- as fast a processor, as big a hard drive, as much memory -- you still wouldn't be getting what you get with the Mini. A comparably priced Windows computer is a cheap Windows computer; a Mac Mini, with its built-in top-of-the-line software, is a digital media appliance that fits on a countertop, connects to your HDTV, stores all your photos, catalogs your music, edits your movies and (if you slap down $100 for the DVD-burning drive) creates your DVDs. Thanks to the Web, it is also now easier to start using the Mac without really going through the hassle of "switching." Key applications -- like e-mail in the form of GMail, or photos with Flickr -- are available on any platform, reducing your dependence on Windows.

Hmm - he needs to take that trip to CompUSA and do a price/feature/performance comparison. Things simply aren't the way he thinks they are. You do get an audience for the Mac - it's the same sort of audience that is willing to fork out extra money for a high end luxury car instead of a Civic or Escort. The Mac has become a luxury item - value buyers simply aren't going to get past the price/feature problem.

I think Apple recognizes this, and is trying to expand their market share horizontally (the iPod) rather than vertically (the Mac). The mini represents Apple's next move that way, I think. I bought 2 ReplayTV's a couple years ago, and forked out good money for them - and what are they? Single purpose computers. The mini isn't a single purpose computer, but it will probably be bought that way - by people who want to deal with photos and movies, and not deal with this kind of nonsense.

With the iLife campaign that Apple is running, it looks like that's what they have in mind.

This is the sort of post that makes the job worthwhile.

Scoble has a post up on why he hasn't hit a specific Tablet PC bug - he shuts down every night and restarts in the morning. Interesting - I used to do that with Windows boxes, but I haven't been doing that in awhile. I haven't rebooted this new laptop since I first set it up; I also left the old one running for days or weeks at a time. For all the crap I give MS, I have to admit - XP is a distinct improvement over their past efforts.

It's still not as impressive as Unix or Linux in that regard - the uptime on my office Linux server is currently 45 days (which is when the last power outage was). The server running this blog? 435 days - we have backup power for that one. It's nice to have stable systems.

We took our daughter to see the movie National Treasure this afternoon. Yes, the premise (that there's an ancient treasure that some of the founders - who were Masons - hidden somewhere in the US after the Knights Templar found it during the Crusades) is silly. The movie is a good action ride though - the pacing is good, and in the whole is entertaining. If you don't go in expecting high art, you'll be just fine. Just take the ride and enjoy it for what it is :)

Periodically, people ask how to go about setting up a VW web application to auto-configure itself at startup. The plan is to make all of that easy via the existing settings tools - as with other VW settings, that stuff could all be loaded at startup. That doesn't hep now though; the existing tools were designed with the idea that you would run a server with an editable (GUI) console and configure, or that you would save the image in the proper state and just run it.

I ran into this same issue when I was setting up the Wiki and this blog server a few years ago, and I created the WaveConfigurator package (check the public store) to deal with it. Basically, it allows you to specify an ini file of settings that will be checked at startup. Here's what the ini looks like for the test server I run on my local Linux box:

[WaveConfiguratorSettings]
type='VisualWave.IPWebListener'
port=2223
baseResolvers=''
baseServices=''
serverNameForSocketBindings='victoria'

The blank settings are a legacy of classic Wave - they allow for the setup of old style servers that use the 'launch' capability. The rest of it is pretty simple - type is the class name, port is the port, and the serverName setting specifies which name to bind the socket to - which is good if you don't want the server socket bound to every IP on the server. I've been using this scheme to have servers start themselves up for awhile now. Combine that with a command line argument to load a specific parcel, and you can set pre-requisites such that a base image will load everything up and just go.

As to how to get that base image - for servers, I don't bother to use RuntimePacker. I simply save the development image headless (i.e., no GUI), and have the startup process configure everything else. The nice thing about that is that I can then script in a command to save the image with a GUI, download it - and debug any odd server issue in the context that it actually occured (try that with a Java or .NET based service :) ).

I've noted before that building and deploying Smalltalk applications is harder than it needs to be - this is my biggest priority over the next little while for Cincom Smalltalk. However, I'm seeing that deployment isn't necessarily a walk in the park elsewhere either. Have a look at Sriram Krishnan's adventures while trying to get a working Parrot.

Scoble points to a post by Claire Giordano about the airport signs in Oakland. I've seen those as well, in both Dayton and BWI. Here's the thing - I'm not convinced that these signs are an improvement over the old "green screen" ones. Why? Simple reason - there was more usable information on the old ones - more actual flights listed per page. With the new ones, I have to spend more time waiting for the monitors to page than I did with the old ones. The new ones might be prettier, but they are actually less useful.

Update: Those signs in the picture are bigger than the ones in BWI and Dayton. The smaller ones I've run across are less useful; the big ones shown in the linked picture look pretty good.

Last week I had the devils own time migrating files from my old laptop to the new one - although, as I stated earlier today, it seems that a large part of the problem was the WiFi card I was using. Still, the whole experience made me interested in considering the Mac. So there we were this evening, at CompUSA - my wife wants a new machine that can better handle video and digital photos than the underpowered system she now has. Yet again, I had it made clear to me that the Mac just isn't in the cards.

Consider what I was looking at - a Mac (a G4 mind you, not even a G5) that ran $999, and a PC that was priced at $800. The Mac had an 80gb HD, 256 MB memory, 2 USB ports and 2 Firewire ports. The PC had 1 gb memory, a 200 gb HD, 7 USB ports, 3 Firewire ports, and a bunch of ports for memory sticks (only a few of which I recognized). The PC came with a monitor (one of the thin screen flat deals) and an HP printer/scanner/copier. The Mac came with a monitor. So heck - a less powerful system, with fewer useful ports - for nearly $200 more.

As it happens, we didn't buy the PC - my wife still wants to shop around. I'd like a Mac, but boy - it would be awfully hard to actually justify buying one with those price points.

The more I read "Collapse", the more I like it. I enjoyed Jared Diamond's previous book, "Guns, Germs, and Steel" as well. That book addressed the question - why did European and Asian societies achieve such higher levels of technology than the rest of the world? To boil it down (read the book!) - it had to do with resource advantages in the areas where civilization arose.

"Collapse" is about something else - why some societies have collapsed (Easter Island, the Maya, the Norse in Greenland, for example) - and why others have succeeded - like the Inuit in Greenland, who displaced an earlier people there and outlasted the Norse. Diamond traces a number of things that led to problems for various societies - climate changes, hostile peoples, loss of trade relations with friendly peoples, and environmental damage. He adds a fifth crucial one - the responses a people have to their changing environmental conditions.

When the term "climate change" comes up, he's not using it to discuss the global warming debate - rather, he's talking about changes in climate that affected a society. For instance - when the Norse settled Greenland, it happened to be in a long term spell of moderate (for Greenland) weather. The seas were navigable (less ice), the growing season was long enough to (barely) support an agricultural society. Things got more difficult during the little ice age, a period when the climate in Europe and the Americas grew much colder. The Norse did not adapt to that change at all well, and died out. The Inuit, on the other hand, lived through that period.

So far, I've read about modern problems in the American midwest (Montana), the disappearance of the Anasazi, the problems of the Polynesian settlements on Easter Island, Henderson Island, and Pitcairn, and about the disappearance of the Norse from Greenland - as opposed to their long term survival on Iceland, and the Inuit success in Greenland.

There's a lot more of this book to read - Diamond addresses the genocide in Rwanda in light of his research elsewhere, and I expect that section of the book to challenge my current assumptions - all learned from press reporting - on what happened there.

There are cautionary tales for modern society in this book, having to do with the growth of interdependence. The Polynesian settlements on Pitcairn and Henderson island were wholly dependent on the "mother country" of Mangareva, for instance - and when that island stopped communicating with the colonies (due to their own near collapse), they both died off completely. I'm not going to get into any of the environmental, trade, or political issues relating to this subject here - it's not something I address on this blog. What I will say is that this book is worth reading regardless of how you see any of those things. Definitely food for thought.

Dave Winer and John Robb are flogging a truly stupid idea - a central repository for RSS subscriptions that would serve as a "solution" to the one click "problem". Go read what Dare Obasanjo has to say about this; he covers the bases pretty well. There are a couple of immediate problems that come to mind:

• privacy concerns - how well protected will these subscriptions be? Are we all willing to show the world everything that we read?
• single point of failure - enough said

The amazing thing is how much traction this dumb idea seems to be getting - witness Tim Bray's non-dismissal of the idea. You know, this isn't hard to solve on the client end - there's the feed:url solution - which many aggregators, including BottomFeeder already support. There's also mime level handling, as per the USM - I will be porting the code for that into BottomFeeder soon, just as soon as I acquire the energy the go back to registry hacking :)

The point is, this is a fairly easy to solve problem at the client side, where the two big problems - privacy and single point problems - disappear, or at least go under the management of the user himself.

Briefly alluded to in Tim's post is another problem, which I think is the ultimate reason that this won't get a lot of traction - where's the business model? What's the motivation for anyone to set up a huge server farm to service all of this?

I let loose with a few rants on the process of transferring files to my new laptop last week, and I figured I should wrap that up. It turns out that at least half of the problem was the network card in my old laptop (or the drivers for it, or something related to that card). The WiFi connectivity with the new Thinkpad (built in Centrino) is much faster. It's clearly not the LAN itself - I haven't changed cable modems or routers. That still doesn't explain why Explorer has to be so much slower than XCOPY, but things would have been a lot smoother with a better network card. So, it's not all Microsoft's fault, and I was a little premature with the level of anger directed their way. I've already shipped the old laptop back to corporate, so suggestions aren't going to help any at this point...

There's a lot more going on in the search space than I had thought. All the talk of desktopsearch tools has covered up some pretty interesting things. The BeyondVC blog has a few interesting tidbits:

• Google is using Answers.com to get the definitions that they are slapping at the top of results pages
• The MyMSN feeds are apparently powered by Moreover, a name I had nearly forgotten

There's more detail at BeyondVC - and aparently, a lot of search stuff happening that isn't about Google or MS.

Tim Bray is worried about security issues with more and more developers including HTML processors in their applications. he points to this:

Atom, like various RSS flavors, lets you include HTML in entries. The IETF requires that all specs have a security section. We were hunting around for a suitable reference on HTML threats and didn't find one. If one exists that covers this modern life, I'd love to know.

It boggles my mind that the W3C (I think this is their problem) or the IETF or *someone* haven't dealt with this. With MSHTML, Gecko, and WebKit, we've started to see many developers incorporate HTML in their applications. The population of apps ready to be burned is growing all the time.

As we all know, standards organizations have limited resources, so perhaps they should hold off on the Modularization of MathML X-Forms over SOAP/BEEP with MTOM base-64 content and take the time to document what's out there now.

Meanwhile, someone's mother is clicking on a popup window that's warned her about "DANGER"...

Of course, I'm completely immune to all of those problems in my aggregator - because the HTML processing and display simply doesn't support any of the things that are a problem - and we aren't embedding anything that does.

When MS was forced by the EU to ship a version of Windows without Media Player, they decided to go with a take no prisoners attitude towards the whole thing - they are going to call that edition "Windows XP Reduced Media Edition." Heh

Update: The EU Commission was not amused by the proposed name

Half the reason IBM moved from VA Java (which was built in Smalltalk) to Eclipse was that it was supposed to allow them to easily support new revs of Java. Well... it looks like that didn't happen - as Bruce Eckel points out. Fascinating

InfoWorld's Paul Krill points out how much has changed in the landscape over the last 10 years:

• Spam was this rather tasty meat product that came in a steel can and had this gook on top of it.
• Steve Jobs was out at Apple.
• Bill Gates was worth only a few billion dollars.
• Sun Microsystems was a company on the rise.
• Hewlett-Packard, Digital Equipment, Compaq and Tandem were separate, competing companies.
• The World Wide Web was not pervasive.
• Linux was Lucy's brother on "Peanuts", with his name misspelled.
• Windows 95 was the most widely anticipated product of the year.
• The National Hockey League locked out its players.
• Phishing is what people did in a small boat on the lake.
• Novell NetWare was a critical product in my area of coverage.

Speaking of Windows 95, remember that rollout, with the Rolling Stones "Start me Up"? Things were different in the Smalltalk world as well - Java was mostly just coffee, and the PPD merger had not happened yet (the announcement was in April of 1995). One other thing that was very, very different - the VCR notwithstanding, TV was still a time-bound thing - "Must See" TV was bound to a specific evening. TiVO and ReplayTV have changed all of that.

If you've sent me mail in the last few hours, I haven't seen it - the mail server I rely on for most of my mail is offline at the moment. If you need to reach me, send mail here.

I was just taking a look at the server logs, and it looks like I'm seeing about 200 downloads of BottomFeeder per day - and almost half of them are for platforms other than Windows. I can say this - Bf wouldn't be nearly as useful as it is if I wasn't getting constant feedback from the user community. Thanks!

I found this article on a blog/media interaction fascinating. The reaction of the newspaper to this incident is a lot like the reaction of CBS to the "Rather-gate" document thing - sheer shock that someone, anyone, would be questioning the high and mighty media. Take this reaction, at the bottom of the article, for instance:

For ombud Parry, both sides should be warned to be careful dealing with the effects of blog-newsprint battles. "I have yet to find anywhere in the mainstream media anyone who really has a handle on bloggers," she asserts. "We are dealing with a relatively new phenomenon."

Over here in the technical realm, we have the technical press, but they seem - for the most part - far, far more savvy to the changed media-scape than the rest of the media. For one thing, there are prominent tech journalists blogging - like Jon Udell and Dan Gillmor. Even the analyst community (no favorite of mine) has a hand in the water - notable is Jupiter Research, which has a bunch of their people out there. Notably absent is the sloppiest of the analyst outfits, Gartner - I guess they are too busy making 20 year predictions to pay attention to anything relevant.

Things have changed. It's not that bloggers are better or more reliable than the media, it's that it's now possible for individual expertise to be applied in ways that it couldn't be before blogs became popular. Reporters tend to be generalists - and generalists often make simple errors in their reporting. What's changed is that those errors can be publicized. A decade ago, no one saw the letter to the editor. Now Google can locate the "what so and so got wrong" story. Time for the major media to take note of that change.

More info from comp.lang.smalltalk on the new Los Angeles area STUG:

Having received several indications of interest in the past couple of days, I think we have enough people to start up a Smalltalk Users Group. We are changing the name from the Pasadena Smalltalk Users Group to the Los Angeles Smalltalk Users Group, as we would like to attract people from the wider LA area.

Toward this end, we are also moving the user group from meetup.com to Yahoo group at

http://groups.yahoo.com/group/lastug/

to simplify communication and coordination between group members. Please join us if you live in the Los Angeles area and are interested in Smalltalk.

I know what Microsoft is trying to accomplish with this policy. Their goal is to have invalid Windows users get the real thing. There's a problem though - by denying patches - including security patches - to invalid users, they are actually punishing the rest of us. How? Well, those systems that don't receive these patches will be more vulnerable to the various worms, viruses, and trojans that are out there. That means that they will be more likely to become unwitting zombies - and thus a hazard to the rest of us.

Now, you might ask why these users are unwitting - don't they know that they have an illegal copy of Windows, pretty much by definition? Maybe not - here's what Ed Foster has on that:

Part of it is the type of user the program is aimed at. According to Microsoft officials, almost one of every four Windows users in the U.S. and other developed countries is using a non-genuine version. And the majority of those don't realize it, because the counterfeit copies are often very hard to tell from the real thing. And they will be the ones who Windows Genuine Advantage tags as having ungenuine software, because those who did consciously get a cracked copy of XP probably aren't going to bother trying to validate their copies.

So we're not talking about people who were trying to rip off Microsoft. Instead, an awful lot of people who paid their money for Windows in good faith are going to discover that somebody along the line - a distributor, a reseller, an OEM -- cheated them. They are just as much victims of the counterfeiters as Microsoft. More actually, because they were in less of a position to defend themselves. Perhaps we should call them Windows' Genuinely Disadvantaged.

Take the PC I'm writing this on - I got it from corporate. I had nothing to do with installing Windows on it, and I'm taking it on faith that our IT group has a valid license. In my case, I'm pretty sure that they do, but I'm also sure that there are end users out there who receive pre-installed systems that are not legal - but were received in good faith. This change by Microsoft hurts the rest of us, because it leaves that population more open to attack

Again, I understand why Microsoft is doing this - I just think that they are letting their legal department run ahead of common sense.

When your Lexus can get sick from a Bluetooth phone, then you know that your car is just too darn smart for its own good. Hat tip to Doc Searls and Kim Cameron:

Lexus cars may be vulnerable to viruses that infect them via mobile phones. Landcruiser 100 models LX470 and LS430 have been discovered with infected operating systems that transfer within a range of 15 feet.

"If infected mobile devices are scary, just thinking about an infected onboard computer..," said Eugene Kaspersky, head of anti-virus research at Russian firm Kaspersky. "We do know that car manufacturers are integrating existing operating systems into their onboard computers (take the Fiat and Microsoft deal, for instance)."

Can I have an analog car, please?

Just like all the children are in the top 50% at school, Joel points out that all of our employees are hired from the top 1% - or at least, we like to tell ourselves that. This is one of the best pieces of advice in the article:

By the way, it's because of this phenomenon 14the fact that many of the great people are never on the job market - that we are so aggressive about hiring summer interns. This may be the last time these kids ever show up on the open market. In fact we hunt down the smart CS students and individually beg them to apply for an internship with us, because if you wait around to see who sends you a resume, you're already missing out.

If you are using VW 7.3 (part of the fall release of Cincom Smalltalk), then you'll wantto bookmark this page - engineering and support will be posting patches there as they become necessary. Visit the main patch page here in order to locate appropriate patches for releases going back to VW 5i.3.

One thing I should point out - those pages do not represent all possible patches for a given release - rather, they are a listing of the most critical ones. If you are having a problem that does not seem to be addressed there, then contact support.

Now that I've got my new laptop set up, I figured I should look at the age of my backups. I have an external (network or USB) 160 gb drive for that, and I decided to move it from a net drive to a usb drive with this machine. That was fine, except that all my files and directories were giving me "permission denied" errors. I had a brief moment of panic, thinking I'd lost my backups - but thenit ocurred to me that I should see who owned the files. Dohh - that was the problem. After I swapped ownership, all was well. Sometimes, the problems are blindingly simple...

Via Phil Windley, we learn the limits of Business Intelligence software:

So I'm waiting in line at Safeway to buy groceries. Like most supermarkets these days, they've got a loyalty card program and offer reasonable discounts at checkout for cardholders. An older man in front of me wants to purchase items that are on a card special, but doesn't have his card and can't remember his phone number. The clerk says, "hmm, wait a minute." He starts punching in phone numbers at random, and after a few tries gets a "hit" and uses it to apply the discount. I wonder who just had hemorrhoid cream added to their past-purchases profile? Gotta love that inevitable weak link in business intelligence systems.

You can see the confusing results of this for yourself by buying books for yourself on Amazon - followed by a few unrelated gift purchases. The recommended reading list then gets to be "interesting", to say the least :)

I pointed to Morgan McLintic's blog earlier, but I should mention that it looks like a good resource on PR stuf. Subscribe here

Travis was wondering where all the innovation was in relation to the Refactoring Browser - and Sascha Doerdelman points to it. Looks like Niall Ross has been busy.

In a post about a lot of little things, Ted Leung points out - in passing - just how truly ubiquitous network access is becoming:

Also, today was the first day that I was actually able to use the Wifi network on the ferry. The WSF isn't quite ready to declare the system ready for use, but the Mobilisa folks let me register for an account, and have answered my tech support queries, so I figure I' m beta testing. So far I've used the network in the Seattle terminal, the Seattle terminal parking area (very important) and on the Tacoma and Wenatchee. Unfortunately the network didn't work in the Bainbridge Island parking area (very bad). And for some reason unknown to me and to SpeakEasy, my DSL provider, I can't access the servers at my house from the ferry network. The Mobilisa folks haven't figured that out either, yet, otherwise I would have posted from the ferry.

Now, put that together with Clemens Vasters' post from yesterday - posted from 36,000 feet via Connexion. We are rapidly approaching the always on world. It wasn't that long ago that there really wasn't a network (for most people, most of the time) - when I started at ParcPlace in 1993, I had dialup access to email, and that was about it. In just 12 years we've gone from that to always on. This is a huge change, and it's one that I don't see a lot of discussion on in the "traditional" media.

I think that these maturity levels that martin Morgan McLintic defines for agency/client relations apply to marketing departments as they relate to business units as well. Lots of good food for thought here

Back in the mid 90's, one of the (ever changing) slogans that ParcPlace-Digitalk adopted was "How fast can you respond?" Interestingly enough, that slogan was somewhat prescient - as Morgan McLintic notes in a post on PR responses to blog attacks:

The challenge from a PR perspective is what to do. To respond formally may exacerbate the problem. To ignore it can cost you dearly (ask Kryptonite). And so here we enter a realm where crisis public relations meets customer service. I think the best starting point is to establish the facts internally - does the attacker have a point? Here you have to be candid and try to see the offended blogger's perspective.

Morgan points out that you have to look at the seriousness of the charge before you decide to respond - I'd add that you have to also look at who's posting it. Like it or not, there are A-listers and there are B-Listers (and people lower down the food chain). There's a huge difference between one of the A listers going negative on you and someone with far less Google juice doing it. Part of your decision making process s going to have to be along the lines of "will searches tuen this rant up?". If the answer is yes, that ratchets the importance of the rant up. Of course, you have to also take the source into account in other ways:

The response needs to be considered carefully. The phrase 'Never pick a fight with someone who buys ink by the barrel' applies here. A blog has a lot of space which has the potential to be filled with a personal diatribe of unfounded tosh. If the response is ill-judged or taken wrongly, the situation could be made worse. As a default, my approach would be to watch but say nothing - crises frequently blow over without turning into a wildfire, despite the initial smoke.

Which is to say, you don't want to elevate the relevance of an attack unnecessarily. This is similar to deciding how (or whether) to respond to a negative review in a trade journal - a lot depends on the source. The decisions don't get easier or harder when the rant is online; they just have to be made more frequently.

Cyrus explains some error reporting code in the current beta of VisualStudio - basically, VS is either going to throw a "recoverable Watson" at you when an exception occurs, or queue them up for later reporting. The idea is to have more information sent back to the team in Redmond. I'd be curious to know how this works out, for a very simple reason - in my experience, people - users and developers both - never read the text of the error messages that come up. They tend to hit <return> and move along. It looks like the default for this is going to be "send the information to MS". It will be interesting to see how that works for them.

Via Blaine Buxton, who got this from lemonodor:

Bruce Eckel, author of Thinking in Java and Thinking in C++, is working on a new edition of Thinking in Java: This language, which was once hailed (admittedly, by the PR flaks at Sun itself) as being "much easier than C++" really isn't anymore.

When your strategy is to add features via complexity, that's where things end up

Hmm - Frank Hayes doesn't sound nearly skeptical enough about Oracle's migration plans:

Let's say Larry Ellison is right. Suppose Oracle does hold on to PeopleSoft and J.D. Edwards users by finding a way to let them migrate gracefully to a best-of-all-worlds merged product that's now code-named Project Fusion and is supposed to be ready by 2008

How will Oracle do that? Nobody knows. It's never been done. Today, ERP migration is hard and painful. Making it easy will require a huge leap forward in migration technology.

But what if Oracle is serious about this and it really happens? Then things get interesting -- and not just for PeopleSoft users.

I think Hayes needs to run the oracle Installer a few times - that'll disabuse of him of the notion that Oracle can pull this off. I still have nightmares about the 8.1.5 installer - half the reason I haven't upgraded the OS on my Linux box recently is fear that I might have to upgrade Oracle - which takes me to this morning's confidence building installation of Oracle client libraries on my new laptop. 600 MB later (for the client?), I had the installation files. Partway through the install, Windows popped up a message telling me that system files had been modified with unrecognized versions... and the installer was the only thing that could have done that. Yeah, that made me confident.

I think it's time for PeopleSoft/JD Edwards users to start sweating - if for no other reason than that they'll have to face the Oracle installation tools...

Bob Congdon points to a snowy tribute to Calvin and Hobbes. I wish we had enough snow here to do some of those!

This is cool news - the Universal subscription client supports a bunch of newsreaders, including BottomFeeder:

I added USM capability for 20 RSS readers into the USM reference implementation client. Click here to download the installer. The installer will prompt you for your preferred RSS reader. The supported RSS readers are My Yahoo!, My MSN, Bloglines, NewsGator Online, Ampheta Desk, Awasu, Bot a Blog, Bottom Feeder, FeedValidator, fyuse, Headline Viewer, IzyNews, mobilerss, NewsIsFree, NewsMonster, nttp://rss, Radio Userland, Syndic8, Wildgrape, WinRSS. I don't spend much time debugging, so please send me feedback and bugs.

There's some interest in a Southern California Smalltalk STUG - see the Usenet post. Apparently, a meetup has been planned. The DC meetup last week went quite well.

Just in case you were happy to see IBM and Sun releasing (some) patents, have a look at Tim Bray's post on what we aren't hearing about. The ones he highlights truly are atrocities...

If you are attending OOP (in Munic, Germany), then check out Joseph Pelrine's talk on agile testing in Smalltalk:

On January 26, I'll be speaking at the annual Smalltalk evening at the OOP conference in Munich. The topic will be on agile testing in Smalltalk

Derek continues to expose poor practices by various and sundry copanies he deals with - today's target is Charter One, and it sounds like they deserve all the scrutiny they can get:

There was an article in the local rag about a theft from the local branch of the Charter One bank. Apparently there was a bag of paperwork and check-deposits waiting for a courier to pick them up. That bag was stolen and absconded with.

According to the article:

A letter from the bank to customers who made deposits that day at the Wall Street office said the bag contained "branch work for the day" and that the depositors must "obtain replacement items for any check you deposited on January 14, 2005."

Boy, that's convenient. Makes me all warm and fuzzy that Charter One handles my mortgage...

I see that there's been another report of paid blogging over in the political realm. I'm not so interested in what the topic was - I'm more interested in the paid blogging itself. I can't find a reference, but a few months ago there was some buzz about a blog equivalent to product placement - i.e., some company pays you $$, and you blog (without revealing the payment) on a product and its virtues. What's going on in politics right now is the same thing. It seems to me that disclosure is the best policy.

If people don't know that you are being paid to promote something - and then that fact leaks out - your credibility is shot. I think this is true in any marketing effort - be it political or business. Once your credibility is shot that way, it's going to be very, very hard to get it back - everything you say is going to be examined with the paid, undisclosed marketing in mind. If, on the other hand, you disclose what you are up to, it's different.

Yes, I live in the suburbs. Bob Congdon notes that his Starbucks density is 19 (5 mile radius). Mine is a paltry 4 :)

Sriram Krishnan attended a TechEd conference, and has a good set of notes posted. I noticed this in passing:

Every attendant gets a number of points, and he can give any number of these points for any feature he believes is important. The features with the most points get implemented. The last time the board took a vote, the edit and continue feature got 0 points, which means it probably won't become implemented in the feature, unless a lot of pressure is applied by developers and clients.

Hmm - I thought I've seen other people (from MS, but I could be mistaken) say that this would be added to "Whidbey". In any event, I find it fascinating that a development board would think this is so irrelevant - really makes a Smalltalker scratch is head and go "huh?"

I see a lot of what Scoble points at in this post, and yes, I'm guilty of this "look at me, I'm so busy!" kind of chest thumping as well (although, at the moment I have only 2 emails sitting in my inbox :) )

All day while going around the Blog Business Summit I've been asking people: "How many emails do you have in your inbox?"
One famous blogger (I won't name names, sorry): "4,200."
Another: "2,100."
Another: "700."
Russell Beattie reported on his blog yesterday: "274."
Me? It's embarrassing how much. 148 in my "priority 0" folder (and a ton more in my lower priority folders).

A slight variation is to "complain" about how many unread RSS items you have. Just in case anyone thought that software developers were really any different than anyone else :)

Via Doc Searls comes a pointer to some funny definitions - I like these:

POCKETS OF RESISTANCE 13 "Sounds like someone having trouble pulling their hands out of their pants pockets." 13 Joe Hutley, Las Vegas, NV.

ENEMY COMBATANT 13 "Makes no sense. Do we have friendly combatants? Neutral combatants? Or how about enemy bystanders? If they are your enemy, just say so." 13 Bill Sellers, Hampton, Va.

Hehe - there are some other good ones there as well.

Slashdot reports that the Enterprise folks aren't giving up on the show - no, they plan some real stupidity to wrap things up:

"It seems Star Trek: Enterprise isn't about to go down without a fight. TrekToday is reporting that Jonathan Frakes and Marina Sirtis will guest-star on the season finale of Star Trek: Enterprise, to reprise their Next Generation roles of William T. Riker and Deanna Troi. Hello stunt casting! The news has been confirmed on Sirtis' official fan site."

Thank goodness for quality SciFi like BattleStar Galactica - where the characters actually seem to have substance, and a third dimension...

The MPAA wants nothing less than to overturn the betamax decision. If they get what they want, it could get very, very ugly...

It certainly looks like ebay will sell anything - you think they could be more careful with their wording? Hat tip to Mike

Panopticon discusses Singletons, mostly (but not exclusively) in the context of VB. I thought this deserved some comment:

In most languages, you have to manage singleton objects yourself. This is usually accomplished by sealing the type (i.e. making it NotInheritable), making it uninstantiable (i.e. making its constructor Private) and exposing a shared property called something like DefaultInstance that returns, well, the default instance. If only the language would do that management for you automatically, how nice would that be? If only...

In Smalltalk, Singletons are typically by convention (there are exceptions; built in things like SmallInteger, for instance) rather than enforced. For instance - say I want a single instance of class MyWebService - I'll typically write code like this:

Smalltalk defineClass: #MyWebService
	superclass: #{Core.Object}
	indexedType: #none
	private: false
	instanceVariableNames: ''
	classInstanceVariableNames: 'default '
	imports: ''
	category: 'Web Services'

"class methods"
new
	^super new initialize.

default
	^default isNil
		ifTrue: [default := self new]
		ifFalse: [default]

reset
	default := nil

The idea is, all usage runs through the #default method. Now, this has pros and cons. The obvious con is that we don't have enforcement of singleton status, merely a convention. You can be a little more rigid by moving the lazy initialization into the #new method - but even then, developers can always invoke #basicNew. The point is, it can be gotten around with some level of effort. Paradoxically, that's also a pro - because it makes ad-hoc testing easy. We can create a new instance "on the side" without munging the "default" instance.

As is always the case in Smalltalk, the decisions are mostly left in the hands of the end developer rather than in the hands of the system. Is that a good thing? I think so, but YMMV

I've been watching 24 for 3 years now (through one absurd plot point after another, but nevermind) - the pacing is what I like better than the stories. Sometimes though, the absurdity of a situation just jumps out and bites me. A couple of examples come to mind from the ast two episodes:

• Jack has to keep the suspect (Kalil) from leaving the gas station long enough to get satellite coverage. He decides to fake a holdup.

Ummm - how about slashing the tires of his car? Or perhaps walking into the store and just being a jerk, forcing the clerk to pay attention to him? That one passed quickly, but the next one was a whopper. The terrorists are using the internet to broadcast a show trial, and the rescue forces are 20 minutes out - so they decide to bomb the place. Ummm - did anyone think to call the power company and just cut the power to the site? They knew where it was - a local blackout would have been easy.

It's things like this that jar you out of the plot and into "wtf??" mode during it. "Enterprise" does that all the time - and 24 is getting worse about it. At least there's Sci-Friday on Sci-Fi channel :)

Ted Leung makes some interesting points about development tools and the power they give you - if those tools actually grok the system they are being used to build

With Python, the tools are different. Right now I use a combination of Emacs and ipython for much of my work. When I start having to interact with code in the application layer of Chandler, then I find it useful to have Wing in order to present a usable debugger interface (I can hear James Robertson cringing all the way from the other side of the country). But none of these tools really understand what a Python program means either.

This is one of the biggest failures those of us in the Smalltalk (and Lisp) communities have - the failure to properly explain to people what it is we have, and why it's of value. Apparently, all most people see is that Smalltalk is different (and to many people, different = bad all by itself). Ted points out the value underneath the hood, and it's something we need to be better about explaining.

So, relative to my various posts on the joys of Windows networking, I now have a question - if I want to copy across the network, why is the DOS XCOPY command so much faster than the Windows shell copy/paste operation? I can't believe that the little paper animation is slowing things down that much; something else is at work here. Very odd.

Ian Bicking has a post on exception handling - specifically, on handling arbitrary exceptions:

But there's other cases. Anytime you won't be there to babysit a process, you need to handle unexpected exceptions. For a command-line utility, you can read the exception when it occurs, no big deal. But for a long-running or batch process you need to intelligently deal with exceptions. Also for processes that are run by non-developers: you want to capture the error information so a developer can look at it, and then try to keep going if you can.

You'll want to read Ian's entire post; I'm only excerpting a small amount here. This does illustrate something of interest to me though - for a deployed application (especially a deployed client application), you want to be able to catch all exceptions. In Smalltalk, that's straightforward - the system already does that.

During Smalltalk development, an unhandled exception will give you a notifier (which offers you the option of debugging). Clearly, that's not what you want at runtime. On the other hand, you would like something like this:

• Catch all exceptions
• If it's possible to continue after handling the problem, do so
• If it's not possible to continue, at least save work and quit cleanly

In a VisualWorks application, the way you accomplish this is to look at class Notifier. What you can do is create a new class - either a subclass, or a separate class that implements the API - and swap it in as the default exception handler. When you package with RuntimePackager, it does this with it's own handler (default behavior - log the error and quit). You can change this behavior to anything you want - on a per exception basis. Very powerful, and extremely useful.