Smalltalk Tidbits, Industry Rants

Who would have thought that my county government was this daring? Normally, they are way too busy installing "Traffic Calming" devices in the road to make things worse. Via Matt Croydon comes word that the local libraries are moving to Linux for their public terminals:

(NewsForge]) Over the past year, the Howard County (Md.) Public Library has migrated more than 200 public PCs from Windows 98 and Windows NT to Linux. These PCs are used both to surf the Internet and to access the library's catalogues. NewsForge recently spoke with Brian Auger, associate director of the library, and the IT team responsible for the migration. We wanted to learn more about why and how it was accomplished, and how pleased they are with the results.

Interesting. I'd guess that licensing fees for things like Office are pretty steep for (typically cash strapped) organizations like public libraries.

It's interesting to watch the way things go with the mainstream development tools. Given a choice, they always seem to go towards higher complexity. Sun could have picked a reasonable model for generics; instead, they got their inspiration from the horror that is C++ templates. It's no better over in CLR-Ville - look at this post from Julia Lerman's Blog - VB developers are looking at the complexity of C#, and having a hard time of it. For all its warts, VB used to be relatively simple and productive. Now, it's a mass of increasing complexity forced by the rigidity of the CLR.

There have been better, more highly productive answers for years - Lisp has been around for decades, as has Smalltalk. People are discovering Python and Ruby as well... and then there's MS and Sun, wedded to complexity.

Dave Astels has posted up SmallMock - Mock Objects for Smalltalk for download. There are versions for VW, VA, and Squeak on the site. The short description:

SmallMock is an open-source tool for creating Mock Objects in Smalltalk quickly and easily. It is similar in approach to EasyMock for Java, and comes in three flavors: VisualWorks 7.1, Squeak 3.4, and VisualAge 5.0.

Mock Objects are especially useful for testing resources whose states are hard to control, or to which access is hard to get. Mock Objects take the place of real objects for the purposes of testing some functionality that interacts with and is dependent on the real objects. There is a site dedicated to the concept: www.mockobjects.com.

This is what Dave was going to talk about at StS 2004, but he had to cancel at the last minute. Check it out and see what it's all about

Update: Posted to the public store - read the license

The rocket scientists at NBC apparently don't want me to watch "must see TV". For the last couple of months, they've been screwing with program start times - moving them from 9 to 8:59, for instance. Why do this? To screw with people who use DVR's to record the shows, obviously. All this really does is irritate a growing proportion of their audience though The more NBC plays around like this, the less likely I am to watch any of their programming - I don't have enough time in the day to second guess them like this. Here's a clue folks: your economic model is broken. We aren't watching your ads anymore. This behavior is simply a last desperate attempt to deny reality - and the net result of it is to tick off viewers who are actively seeking your programming. Is this really who you want to cheese off?

Mark Pilgrim covers the MT imbroglio pretty well - but draws the wrong conclusion, to my mind:

It's not about who has a right to make a living (everyone does); it's not about how nice Ben and Mena are (I've met them, they are very nice); and it's certainly not about eating. I've taken the $535 that Movable Type would have cost me, and I've donated it to the WordPress developers.

It's not about money; it's about freedom.

I don't think "freedom" is the operative issue here, exactly. This is a much older problem with respect to a product - appropriate price points. After reading the many, many takes on this (including comments to yesterday's post), I've come to the conclusion that Six Apart probably (it's early yet; I could be wrong) misunderstand the market they are in. Their pricing indicates that they want to play at the low end of the Enterprise market - but that doesn't seem to be who's using their product right now. The problem is, by putting out their product at a low (free) price, they made it very, very difficult to move up the price chain as rapidly as they are trying to.

I have a lot of sympathy for this problem - as a Product Manager, I am involved in discussions over the pricing of our product all the time - and believe you me, it's not simple. You have to pick a price that allows you to pay your ongoing costs, or you don't stay in business. You may have costs that are too high, or prices that are too high - or the reverse. Screw this up and you limit your ability to grow, or grow too quickly without bringing in sufficient revenue. It's not an easy balance, and I think Six Apart is only just walking into this particular buzzsaw.

Ralph Johnson pointed me to this page - the projects he had students working on at UIUC this last semester. Some of these are really cool sounding. A couple stood out to me:

• An auction/agent system in Smalltalk
• CURSES style UI in Smalltalk
• Open SSL Tools
• An FTP Server

Also lots of games - check them out. maybe I'll make one of them a BottomFeeder plugin...

Sci Fi Wire reports that we may see an Ash vs. Freddy and Jason movie - but it will be hard to bring off:

Bruce Campbell, who played Ash in the Evil Dead horror films, told SCI FI Wire that there is "some validity" to the rampant rumors about the possibility of a movie that would pit his character against Freddy of A Nightmare on Elm Street and Jason from the Friday the 13th franchise.....
Campbell added, "[That] means you have 17 lawyers, each going, 'Yeah, look at my franchise. Yeah, look at my franchise.' So you have to get past that. Then you have to get past the question of 'how will the characters be treated?' What would you do with Ash? There's no way I'd be in it if I lost. No way. The good guy has to win, especially in that movie. You couldn't kill two whole franchises, but we could sure maim them."

Now that would be a cool flick! Hail to the King, baby

Here's a press release STIC just dropped:

Seattle Hosts Successful Smalltalk Solutions

Cary, North Carolina, May 14, 2004 -- The Smalltalk Industry Council (STIC) is happy to announce that this year's Smalltalk Solutions in beautiful Seattle was a rousing success. Smalltalkers of all ages and abilities from throughout the world descended upon the Seattle conference to experience the rich variety of talks and presentations that were offered. The hotel buzzed with conversation as colleagues both old and new traded ideas, shared their experiences and debated throughout the three-day conference. The Council would like to thank all the presenters and attendees who helped make the show one of the best ever.

The presentations, tutorials, and BOF's were a big hit and offered a little something for everyone at the show. Attendees selected from a wealth of choices from pension-plan management and emergency-services dispatch to internet syndication and garbage collection. The talks generated much interest among the conference-goers with many conversations making their way out of the classrooms to the exhibit hall where others could enjoy and join in.

It was standing room only when the conference keynote speakers took the floor. George Bosworth of Microsoft discussed Smalltalk and CLR virtual machines. Former Hotspot architect Lars Bak discussed Smalltalk for very small embedded devices. And Seaside inventor Avi Bryant talked about Smalltalk's role in redefining web development. STIC would like to publicly thank the keynote speakers for the wonderful job they did in informing and enlightening the conference on such exciting topics.

About STIC

The Smalltalk Industry Council (STIC) is a nonprofit trade association whose goal is to promote awareness and increase demand for Smalltalk. STIC was reorganized in 2002 by Cincom, GemStone, IBM and Knowledge Systems Corporation, creating a cohesive Smalltalk community where information, technical issues, new ideas and concerns are openly discussed to benefit businesses as well as the software industry. STIC's membership consists of users, service providers and vendors of Smalltalk tools, components, databases and services. For more information on STIC, please visit www.stic.org

Contact:
Jason Jones
Smalltalk Industry Council
919.789.8549 x.21

Stay tuned for news about next year's show - if you have suggestions for a location, let us know now

Here's an interesting post from Dare Obsanjo on XML tech. He's explaining why they don't plan to support XSLT 2.0, XPath 2.0 and XQuery 1.0. Here's the basic explanation:

We'd rather implement one more language that we push people to learn than have to support and explain three more XML query and transformation languages, in addition to XPath 1.0 & XSLT 1.0 which already exist in the .NET Framework. Having our customers and support people have to deal with the complexity of 3 sophisticated XML query languages two of which are look similar but behave quite differently in the case of XPath 2.0 and XQuery seemed to us not to be that beneficial.

This all makes sense - you have to make choices about what you will and won't support - you can't support everything out there because there's simply not enough time/resources to do so. What struck me was this:

XPath 2.0 has different semantics from XQuery, XQuery is strongly and statically typed while XPath 2.0 is weakly and dynamically typed. So it isn't simply a case that if you implementing XQuery means that you can simply flip some flag and disable a feature or two to turn it into an XPath 2.0 implementation. However all of the use cases satisfied by XPath 2.0 can be satisfied by XQuery

Sadly, this is still how most of the software world thinks - they pick rigidity over flexibility, and static over dynamic. Productivity goes out the window in the cosmic quest for "safety" (which is illusory - the Capers Jones data from the 90's shows that error rates are lower in languages like Smalltalk than they are in languages like Java) - and yet, the major players keep marching down the path towards rigidity. MS is doing it universally - VB, which used to be somewhat dynamic is being shackled to the rigidity of the CLR, and is losing any residual goodness it had left.

All the progress in software is coming from the margins at the moment - people experimenting with Python, Ruby, Smalltalk, Lisp, and other dynamic systems.....

Dr. Gregory V. Wilson wants more extensible programming languages, and points to Lisp as an example of one. The Planet Lisp crew picked up on that here. If you read the paper, you'll see that Wilson wants simplicity and extensibility. Well, another possible answer to that is Smalltalk - it's a very simple, very powerful language that lets you "extend the language" in very simple ways - take a look at the implementation of #select: in Collection, and then ponder the fact that you can add your own additions to class Collection as needed. None of these things are going to happen in C# or Java - they are both too rigid. VB has become rigid by forcing it onto the CLR. We need something better, but it's not going to come from the mainstream....

Don Park links to an interesting post from one of the commenters on Six Apart's new licensing for MT. It's kind of amusing to watch. I'm sure that a lot of people have enjoyed using MT for free, and gotten a lot of blogs set up. Now ask yourself - what has Six Apart gotten for that? And don't say "positive press" or "love from the open source community" - because neither of those things pays any bills. There's a simple reason that MT is going to start costing real money - it's a business, and I'm sure that the developers want to be able to eat.

Not everyone wants to give away their software and "sell services around it". That often requires travel, and guess what? Not everyone wants to travel. Take it from someone who's spent a lot of time on the road - you pay a price for it. You don't expect your stereo to be free, or the TV - why the heck should you expect software to be free?

John McIntosh's StS 2004 notes start here on the Camp Smalltalk Wiki - links to successive reports are at the bottom of each entry. I am still expecting more notes, and I'll post them or link to them as I get them. Enjoy!

Google has added Atom Feeds for newsgroups - for example, here's the feed for comp.lang.smalltalk. This adds fine in BottomFeeder, so long as you are using the dev stream (where 0.3 formatted feeds are fully supported). The feeds aren't that useful though - only partial content for each item is included. It's not time to abandon my Agent yet....

The W3C intends to get involved with Atom. If you thought there were bizarre happenings there before (mime encoded, large binary data in the feeds!) - just wait until the wizards at the W3C get done with it. At least the current specs are written in identifiable English - the W3C positively strives for odd wording and hard to understand sentence structure....

Yesterday, I had a request come through the Smalltalk IRC channel - why not allow users to eliminate the horizontal scrollbar on the grid view in BottomFeeder? Well, that was a reasonable request - it's now a setting in the latest dev stream build. The way it's implemented illustrates perfectly why the Pollock effort is so important. Take a look at the code that turns the scrollbar off:

| sub ds |
sub := self widgetAt: #feedID.
ds := self
     getComponentFromSubcanvas: sub
     withID: #Dataset1.
ds component container component component components first container useHorizontalScrollBar.

Yes, it's that ugly. The issue is the way the VW GUI adds "decorations" like scrollbars - they are added in via wrappers. It's not simple to navigate down to the correct layer of wrappers - I figured out how to get there via an inspector on the running application. Now, once you get there, there's a nice simple API to turn scrollbars on/off - and yes, I suppose one could surface that via a higher level APi that hid all this nuttiness (but that's non-trivial - wrappers get added based on a lot of variables). This is why Pollock is the future - no one should have to do this to toggle scrollbars :) Thanks to Travis for pointing me in the right direction when I was taking a look at this last night....

There's a Seaside Tutorial for the VW port of Seaside - grab vwnc, then get public store access - and have a look.

Patrick Logan is not impressed with the Tablet PC:

The software sucks. Badly. Really, the pen interface seems slapped on top of the typical mouse-based Windows. And the total experience is *horribly* schizophrenic. The pen/tablet experience is trying to get along with its mouse-based alter ego, but they just seem uncomfortable with each other. Their like relatives who see each other once every year or so. They have something in common, but no deep understanding of each other.

Hmm. This is a slightly less rave review than I've seen from the Tablet evangelists. I'll take a look eventually, but:

• Even if the software is flawless
• And the price differential is only $500 (as Scoble says it will be)

I'm not going to shell out $500 extra for the ability to write slowly on a small screen. I don't touch type, and I still type faster than I write. This is looking a lot like a solution desperately in search of a problem to solve....

Scoble raves about the Tablet again:

But, this whole thing points out a disconnect. I'm using the latest latest stuff. It has radically changed my life. I can't, and will not, buy a portable computer without the Tablet PC features anymore. It's changed my life that much.

Why? Well why don't you meet me in Southwest Airlines on Friday evening and I'll show you why the Tablet PC is so superior for air travelers?

All this reminds me of 1986. Remember that year? Remember what they were saying about the Macintosh? "It's dead," they said.

Notice one crucial thing about the Mac - the market share. And it's not because DOS was better. It's because a Mac was $2k - $3k more expensive than a DOS box was. That's why I bought a 386 DOS box instead of a Mac back in 1987 - I was unwilling to part with the extra $2K. Well, the same now applies to Tablets. Ink is nice and all, but it's not $2k of nice. It's a niche product until the price point equalizes.

Scott Johnson points to a really, really nasty attck based on presumed trust - and he's dead on about how easily people (tech geeks included) will fall for this. That is one nasty, nasty idea there.

TechRepublic.com reports on a judicial block of a SpamCop blacklisting:

A Northern California District Court judge issued a temporary restraining order to prevent SpamCop, an antispam operation, from interfering with messages sent by alleged junk e-mailer OptInRealBig.com.

Now, spam irritates me - a lot. On the other hand, I despise blacklists. The way SpamCop runs them, it's virtually impossible to get removed from such a list, and they occasionally block mail from large ISP's - I've had them block outgoing mail from my ISP's email server. A pox on both their houses....

We have a new blogger on the Cincom Smalltalk Blog site - Professor Ralph Johnson of UIUC (and one of the co-authors of the GOF book). The blog is here - subscribe to his feed here.

There are a few posts on the Tablet PC this am - Scoble over here, Julia Lerman here, and the story that spawned these posts here. I think the lot of them miss something crucial - IMHO, the Tablet is a niche product. Yes, I'm sure that people do like them when they see them. I'm also sure that most people have no real need to scribble on a screen. Think about it - how many tasks do you actually do in a day that would justify the extra cost of a Tablet? To my mind, the Tablet is a great product for a set of vertical markets - and a nice to have, but not important thing for everyone else.

Now, that may change - 5 years ago, notebooks were not something that most people thought of as a desktop replacement - that's not the case now. Over time, the Tablet may become that way as well - but I think this will happen only as the price differential drops towards nil.

Update: Matt Crowden makes my point:

Price is probably the number one factor that is keeping the Tablet PC install base down.  I've looked at a few TabletPC models at my local Micro Center, which usually has a few in stock, and their sales guys know much more than your standard CompUSA or Best Buy.  I played with a Tablet PC.  It was a compelling experience.  Unfortunately it wasn't two grand compelling.  It wasn't twice as much as I ended up paying for a higher spec'd but bigger (and non-tablet) laptop that I ended up getting

Price is the reason I bought a 386 based PC back in the late 80's instead of a Mac. I liked the Mac - but I didn't like the $2k price differential. Ditto for the Tablet...

Patrick Logan points to a ComputerWorld case study of the Kapital system at JP Morgan:

Technology is essential to the success of any high volume derivative business. The technological underpinning of Kapital is based on one of the earliest and some would say most pure object oriented development languages currently in existence 13 Smalltalk. The current incarnation of this language called VisualWorks has been married with a database technology called Gemstone. The combination of Smalltalk and Gemstone was the technology of choice at the start of Kapital. It continues to be the technology of choice today as it enables the following system characteristics:

• The creation of an application architecture that hides the technical complexities from financial developers.
• Distributed calculations are decoupled from the financial domain and increased compute power is achieved through machine on-boarding (e.g. recent machine on-boarding using PC blades and Linux)
• Allows business users and financial developers to create any type of new business object and save it and share it.
• Allows for disparate development teams, who are focused on different business areas, to develop solutions at different rates

Why are they using Smalltalk with all the Java and .NET buzz about?

Identifying new market opportunities is available to all investment banks. However, profitability and success come from being able to execute on those new market opportunities before the competition. Using Kapital, enables both business and IT staff to prototype and develop new ideas that can be promoted into product sets. This rapid turnaround of idea to product empowers JP Morgan to market and trade new deals before most of the competition. In the world of highly complex derivative products, JP Morgan is the number one investment bank.

Oh, and that canard about it being "too hard to find Smalltalkers?"

With the size and complexity of the system (14,000 classes and 400,000 functions) written in an esoteric language and database, one would assume that the technology would be the limiting factor in its continued growth. However, the greatest difficulty for Kapital is not the technology, but understanding the business and business processes themselves. With the developer firmly in mind when Kapital was originally put together, new financial developers can be on-boarded and be productive within four to six months

Go read the whole thing - and then download Smalltalk and try it out.

Every so often there's a "where are all the Java apps" thread on the comp.lang.java.advocacy newsgroup - for instance, here's the latest thread. Well, that got me to thinking - RSS/Atom/Syndication has gotten to be a fairly active application area - there are many, many readers out there - here's one list of the most popular ones. What I've noticed is that there are a bunch of .NET based aggregators, a bunch of aggregators specifically for Windows and/or Mac - but none of the commonly used aggregators are Java based. Now, I know that there are lots of Java developers out there - and yet somehow, no one has created a widely used aggregator in Java. That's fascinating. Oh, yeah - there is a widely used Smalltalk based aggregator :)

It's a good idea to consider every post you make to a blog - and Mark Pilgrim points out today's example: Google re-edits. I seem to recall pointing this out awhile back....

Staffen Malmgren discovers Seaside, and likes what he sees:

At first look, Seaside doesn't look much different than ASP.NET. It's all about modelling your application's interface in terms of objects and methods ("messages" for you Smalltalkers). Pages are built up of components, inheriting from System.Web.UI.Control (ASP.Net) or WAComponent (Seaside) that can include other components. When the user does things with any component, it results in events being fired (ASP.NET) or messages being sent (Seaside). Both frameworks seem to strive to abstract away the request/response nature of the web, and to allow the programmer to use a more event-driven approach to developement. In addition, seaside uses at it uses continuations to make it possible to, for example, ask the user something (similar to how a modal dialog would do it in a normal GUI enviroment), and then do something with the provided answer -- all within the context of a method.

The main difference is that programming in the Seaside framework results in a lot less housekeeping code. The object is really a ordinary object, except that it's executed through the web. The difference didn't really dawn on me before I tried to recreate WACounter as a ASP.Net Server component -- it did involve a whole lot of code to handle events, manage viewstate and so on.

It's cool that people are finding out about Seaside and what it can do.

The Baltimore Sun answers the questions you didn't know you had.

So Scoble says that the XP SP 2 rev will have IE mods. Will this fix the absolutely horrible thing that the last "fix" did to IE? Apparently, there's a security hole in IE involving the use of urls of the form: http://username:password@www.host.com. You would think MS would fix that, hmm? Why no - their solution in the last set of patches I grabbed from the Windows Update (read: downgrade) site just turned that feature off. You try to use it, and IE gives you a "Syntax Error page". Grrr - at least this explains why I've been getting mail from people trying to download CST NC via http. Here's what Scoble says about the fixes:

While we're on it, this page has one falsehood. It says that the next version of Internet Explorer won't come out until Longhorn. That is absolutely NOT true. The next version of Internet Explorer comes with a ton of security fixes, and a pop-up-ad blocker. It will be included in Windows XP, Service Pack 2. For free.

It darn well ought to be free, when it's a downgrade. There should have been a big warning on the last set of patches: "Warning - this update turns off useful features of IE, because we can't be bothered to fix the actual problem. Download at your own risk".

Bob Congdon points to this excellent essay that I've been meaning to discuss:

The golden rule of programming has always been that clarity and correctness matter much more than the utmost speed. Very few people will argue with that. And yet do we really believe it? If we did, then 99% of all programs would be written in something like Python. Or Erlang. Even traditional disclaimers such as "except for video games, which need to stay close to the machine level" usually don't hold water any more. After all, who ever thought you could use an interpreted, functional language to decode Targa images, especially without any performance concerns?

Follow the second link to the original essay - it makes a point about how fast things now perform even in fully interpreted languages on modern hardware - pointing to the absurdity of things like primitive data types in Java. In the vast majority of applications now, that counts as premature optimization. The software industry is guided more by inertia than logic. People continue to select unproductive languages based on the "everyone else uses it" mindset. They stick with manifest typing, even when the best data we have (from Capers Jones) - shows that languages like Smalltalk produce lower error rates than languages like Java. This in turn empowers the managers who want to offshore development - if the results are bad in any case, they may as well be bad for less money.

I'd say that the software industry is where manufacturing was in the 70's - complacent and ready for a kick from foreign competition - but I don't see that happening anytime soon. Why? The big Indian firms are all into huge masses of paper and CMM compliance levels - my cantankerous theory is that you should avoid any firm that claims CMM-5 like the plague (unless you like rigid requirements and huge backwashes of paper). If you really want to outpace the competition, you might want to pick up a dynamic language - Smalltalk, Python, Ruby, Lisp - and run with that.

Bad news for those of us who like coherence in a series - "Enterprise" looks to be back next year, with Berman still running things. If UPN wants to improve the series, the best thing they could do is to fire Berman. Can you imagine a Star Trek with Joss Whedon - i.e., someone who can actually create a story arc that hangs together? The best I can say for "Enterprise" this year is that - unlike "24" - it hasn't completely jumped the shark....

I posted on this topic back here and here. Now, I've come across this perspective on the issue from a woman "in the trenches" as it were. I can add a side note to this - I introduced my daughter to this Girl Scouts site, and she explored it for a bit - but peer pressure intervened and she's back on NeoPets - because all of her friends are there. Now, there's some (small) value here - she's starting to learn bits of HTML and CSS - the site allows you to build pages - but it's not exactly what I had in mind, either. I'm not sure how - as a parent - you manage to redirect towards "better" content. I mean, it's not as if NeoPets is a bad site. Time wasting, sure (but then again, no more so than BSW, where I spend inordinate amounts of time playing Puerto Rico....)

As much as I think Atom is a waste of time, end users shouldn't have to care - that's why BottomFeeder supports the format (both 0.2 and 0.3 - change your upgrade path to 'dev' in order to get proper support for 0.3). So I was a little surprised to see Scoble complain that Radio still doesn't support the format. That's just silly. Atom is a clone of RSS with different tag names; adding support to a tool that already handles RSS is trivial (it took me less than an hour for Bf). As I recall, I only needed a couple of slight additions to the domain model (the additional dates, for instance) - almost everything was a simple mapping issue....

In a Planet Lisp entry, I see that this blog shows up in a PubSub search feed for Lisp references. Heh. I mention Lisp from time to time when I'm discussing the general topic of dynamic languages, but Planet Lisp is where you want to go for Lisp Info. Also Bill Clementson's blog.

MSNBC's business wire picks up the Cincom Press Release congratulating Alan Kay.

James Gosling says that the world needs more "crazy" people:

There's an interesting article at Wired: NASA Funds Sci-Fi Technology. It's all about a part of NASA that funds projects that many would consider completely crazy....

Most real innovation is done by crazy people doing crazy things

Maybe he could follow that advice, and advocate for a JVM that would properly support dynamic languages. Nah, that would be crazy....

We had one heck of a set of thunderstorms roll through the area last night - at one point, my wife and I ran to the door to see what had gotten hit (a particularly loud clap of thunder that seemed simultaneous with the lightning. I'm sure it hit something; I haven't had cable modem service since then :) The cable TV signal has been horrible as well; I suspect that the signal reaching my house is too weak to get me net access. So here I am, feeling the power of dialup. Maybe I should have held on to that AOL CD that came in the mail Saturday....

After many requests, I've added a preview function to the comment form on the blogs here. It only works if you have cookies enabled though. There's also a pointer to the wiki style markup rules that the comment form supports - if you check the appropriate box, those rules will be used when you post.

The NYC Smalltalk User Group is running a poll (membership required) - they want to know whether an StS in New York City would pull in people. Our (the STIC) primary concern with an StS in New York has been the high cost of New York City. In any case, if this is of interest to you, join the STUG's mailing list and let them know what you think

Some image formats are "lossy" - i.e., putting an image in that format can lead to lost information. I'm starting to feel like Java and the CLR are "lossy" with respect to the entire field of software development.

First, there's this post from Roy Osherove. Now, Roy is a good guy - I've had some very pleasant exchanges with him via the comment sections of our blogs. But look at this - in response to a post I made about BottomFeeder plugins:

I wonder how plugins are supported in SmallTalk. What's the Smalltalk equivalant of Reflection, for example? of a DLL? dynamic loading?

Is there an equivalent of reflection.... It's kind of hard to know where to start with that, honestly. The reflection that actually exists in the CLR is a pale reflection of what you get in Smalltalk. I mean, take a look here - that's a paper from OOPSLA '89! I'm not quite sure how you would implement Method Wrappers (the cited page is old, and newer code is already shipped with VisualWorks) in C# or Java. Reflection in Smalltalk is just so much more powerful than what is available in the mainstream languages, but the mainstream languages are so prevalent that it's as if the information has been lost.

Take another example - this post by Victor Ng - he asks which camp is worse, Lisp or Smalltalk:

Which camp is worse? The Smalltalk people or the LISP people? It seems that both those groups just love to say "we did 20 (or 50) years ago"

I suppose it does come off as arrogant and irksome - but it's the simple truth with respect to so many things in software development. Just take a look at one of the big things buzzing around the software world at the moment - self healing systems. In Smalltalk (or Lisp, or other similar systems), new code can be loaded into a running system and override existing methods - or add new methods to existing classes - or add attributes to existing instances - all without having to take the system down or having to worry about the state of objects in memory. Heck, I patch this blog server that way all the time. Lars Bak talked about this in the context of small devices, explaining why Smalltalk was such a better choice - the ability to do on the fly updates.

The dead hand of rigid systems holds the entire industry back. You have to jump through enormous hoops just to accomplish the simplest thing. I'm hardly the only one thinking this - I referred to this essay just the other day. Developers new to Smalltalk tend to pick this up right off

We need to do a better job of explaining this - or else the information loss is just going to continue