Smalltalk Tidbits, Industry Rants

Andrew Binstock has written some interesting rubbish on library vs. language feature:

If one is to believe that threads will be on nearly all desktops soon 14and I think this is now a conservative projection 14then threads processing should be part of the language, rather than a series of function calls.

Who cares? That isn't the problem when writing threaded code; the problem is making sure that code is thread-safe. Here's Smalltalk code that forks off a (lightweight) Smalltalk thread:

[self doComplexWork] fork

Ok, is #fork a library call or part of the language? Why do you care? More importantly, is it threadsafe?. Then there's this:

Another conservative assumption, in my opinion, is that XML processing will become a primary activity of most software. So will accessing resources across the network, and then combining network access and XML into Web services. If so, then Java syntax would benefit from reflecting these features natively in the language. It seems reasonable to assume that URLs 14or, more accurately, URIs 14should be a native data type. And why not native XML types 14although this would be more difficult.

That's just what the curly brace crowd needs - more data types to make learning the language harder. 50 reserved words aren't enough for some people, I guess. Here's some Smalltalk to parse XML: XMLParser new parse: someXMLText. Binstock wants what, exactly - a set of DOM types to screw up Java and C# worse? When was the last time he was near code, and why would anyone let him near any? Sheesh...

Michael Gartenberg thinks that the delay of LongHorn (2006 at least, quite possibly 2007 or even 2008) is a sign of trouble for MS. His evidence?

Microsoft touts Longhorn as revolutionary and says it will make Windows XP look as pale as Windows XP made Windows 98 look. All well and good, except for one thing: A good deal of the market never made the leap to Windows XP. That's amazing when you consider that Windows XP is probably the best operating system Microsoft has ever released, whereas Windows 98 was one of the worst. That stall in the market -- a large number of customers holding on to old operating systems such as Windows 95, 98, NT and 2000 -- is combining with the delays for Longhorn to put Microsoft at potential risk.

A stalled user base is perilous, especially when users are sticking with a product as poor as Windows 98. That means they're saddled with lousy performance, unreliable systems and unsecured ones as well. The second troubled front that Microsoft faces concerns a market that's starting to look for alternatives.

Now admittedly, inertia accounts for a lot of this - but still, his point that the stall on Win 98 is a marketing failure is a good one. XP is a big improvement over 98, and there's really no good reason to stick with it - and yet people are. This does tend to point to a marketing failure at MS. In Gartenberg's words:

It's not about bad product, but rather poor marketing and evangelism, the third troubled front. Let's face it: If you can't show the market value of Windows XP over prior efforts, you're not doing an effective marketing job.

Ouch. Can't say I disagree with him, either. Over to you Scoble :)

I'm in the midst of making BottomFeeder internationalization aware; that means changing all the string refs into UserMessage objects. This is somewhat tedious, but not hard. What I'll be looking for is volunteers to do translations of the catalog file. Now, don't everyone jump up at once :)

If you want to know how BottomFeeder went from concept to deployment, then you'll want to attend my case study at StS 2004. Register today!:

BottomFeeder - a Smalltalk Development case study
presentation
James Robertson: Cincom
Wednesday 8:30:00 am to 9:15:00 am

Abstract: I'll discuss the process of building, deploying and updating a VisualWorks Smalltalk application. The focus will be on BottomFeeder - http://www.cincomsmalltalk.com/BottomFeeder - an open source RSS/Atom news aggregator. This should cover

• Development issues (dealing with RSS, character encoding issues)
• Deployment issues (building a deployable and installable application)
• Cross platform issues (Unicode, character sets, libraries, look and feel)
• On the fly updating of a deployed application

Bio: I got started in Smalltalk quite by accident in 1993 - I was in between consulting assignments at Booz- Allen, my employer at the time. Booz Allen had a training contract with ParcPlace, but had lost both of their instructors. I got picked because I had some teaching experience - 2 1/2 years of junior high and high school. They put a junior guy with no training experience, but some (about a year) Smalltalk experience, figuring that the two of us would figure it out.

I spent 9 months teaching for Booz Allen, but got lured over to ParcPlace - I decided that I would much rather be where Smalltalk was being created! I spent almost two years teaching the intro class before I moved into sales - as a sales engineer. That got very hairy over the next 4 years during the PPD nightmare and the ObjectShare confusion.

When Cincom took over VisualWorks in 1999, I came along, retaining my role as a sales engineer. After about a year, I moved up to Product Management, which is where I still am.

See you in Seattle!

If you are looking for Smalltalk work in Europe, I have a lead on positions in Switzerland - you need the ability to work in Switzerland quickly - a Swiss passport, EU passport, or anything else that would clear you to work in Switzerland easily. Interested? Contact me; the person interested doesn't want their email address posted, but I can put you in contact with the opportunity.

Just look at what you need to do in Java to get a runtime view of objects. Gee, in BottomFeeder - including the runtime - I can just pop an inspector. It's so nice to have real objects at hand instead of pale shadows...

Scoble reports that those horrid Office 2003 ads are toast. As he says, perhaps MS could actually give me a reason to care that Office 2003 is out?

Awhile back, Eric Raymond laid out the pain involved in setting up a remote printer using Linux. I came across this post today, which is a response to that, and to the general problem - up to the quoted section below, it's quite reasonable:

But the whole A.T. angle is quite disingenuous. It wasn't A.T. who couldn't connect to a shared printer. It was Raymond himself who couldn't figure it out. Yes, I see the point that if it were so easy and obvious that A.T. could do it, a nerd like Raymond could do it too. But this is putting the horse way in front of the carriage. In what world does the "archetypal nontechnical user" have two computers connected by Ethernet? When A.T. needs to configure a printer, it's going to be connected directly to her computer, not shared over a network.

Hmmm - I don't think this guy has seen real users up close either. My daughter is in a girl scout troop, and I was talking to the co-leader of the troop the other day. They have a home PC, and a notebook the husband uses at work. They recently signed up for cable modem service, and got Comcast to set up a WiFi router for them - this is a standard Comcast service now. These people are not geeks - and yet here they are, multiple PC's set up on a LAN, with a router, and shared printers in the mix. Here's a cluestick - this is becoming very, very common - as I said above, Comcast now offers to set up a WiFi router as a part of their standard installation. The upshot is, it's worse than you think:

Furthermore, the "I thought I was the only one" response begs the question: what planet are these guys from? Isn't it common knowledge that desktop Linux usability tends to suck? How can anyone write an essay proposing to fix this without mentioning, let alone responding to, Matthew Thomas's seminal essay, "Why Free Software usability tends to suck"?

It's not just Linux that baffles people. Go ask around about anbti-virus software, or firewalls - you'll get a pile of "huh, what?" responses. I had to install my neighbor's printer (not a network printer even!) on Windows, because they had no idea how. Linux has usability issues? heck, PC's have usability issues. Even Macs are too hard for most people. Don't believe me? See all those VCR's with flashing 12:00 clocks? There's your evidence - people want an appliance that just works, not a fulltime hobby that they need to tweak day and night. Linux, Windows, and yes, Macs - are all too hard for most people's tastes. The entire industry has quite a ways to go before we get to real "ease of use" as most consumers understand the term.

On an amusig side note, there's a fair bit of irony here:

If there's a glib, nutshell synopsis for why Linux desktop software tends to suck, it's this: Raymond and his ilk have no respect for anyone but themselves.

They have no respect for the fact that UI design is a special talent.

They have no respect for the fact the good UI design requires a tremendous amount of time and effort.

Heh. This from a guy who's site is dark gray with white text. Sheesh. Could you make it harder on my eyes if you tried???. Now, having said all that, there's a really good point down towards the bottom of the post:

It's pretty hard to sell "services and support" for software that fits that bill. The model that actually works is selling the software itself. This is politically distasteful to open source zealots, but it's true - and it explains the poor state of usability in open source software.

Raymond also complains about CUPS's shoddy and inaccurate documentation, but that's just another side of the same glove. Technical documentation is also hard work, and requires talent to be done well. Writers need paychecks, too. (Trust me.)

Very, very true - and it's what Clemens Vasters was getting at here - a post that was profoundly misunderstood - see this response by Ryan Lowe, for example. There's an old adage that says a lot here: You get what you pay for. If you aren't willing to pay, you are far, far less likely to get useful things out the back end....

There has been a bug in the NC download application for a while now - if the email the server sent failed to go, then it was just lost. I fixed that this morning - the server is now logging such errors, and running a background process to attempt resends. That should clear up one of the bigger issues with the download application. There's another issue with bad error messages, which I intend to take a look at this afternoon. Thanks for your patience!

Update: - as it turns out, one of the larger issues was error reporting - the application would tell the end user "username is already in use" regardless of why the username was invalid - very confusing, and a big error on my part. I've fixed that, so the application should be clearer about issues now.

Interested in Embedded Systems work in Smalltalk? Then don't miss Maurice Rabb's talk on the topic! Register today to check it out:

Microlingua
presentation
Maurice Rabb: Katalytyk, LLC
Wednesday 8:30:00 am to 9:15:00 am

Abstract: Dynamic object-oriented programming languages are rarely used in small or real-time embedded systems. The presentation will discuss implementation techniques originally developed to enable Smalltalk to run efficiently on tiny consumer electronic products such as animated watches. The work has application in other domains including wireless, medical, industrial, and military equipment.

Topics will include developments since Smalltalk Solutions 2003. One of Smalltalk's most cherished aspects is its expressive power. Microlingua introduces a fast and consistent conceptual model for handling numerics, im|mutability, execution context, and concurrency. It improves the safety and power in expressing these fundamental concepts.

Bio: Maurice Rabb is the principal of Katalytyk, LLC in Chicago. He specializes in new product development, conceptual blockbusting, and the creative application of technology -- in short, acting as a human catalyst. Mr. Rabb earned his BS and MS in engineering from Stanford University.

See you in Seattle!

So here's the blog that Scoble was absolutely gushing over the other day. It looks like what Scoble described - a bunch of posts that originated as email to Scoble. Hmmm. Exactly how is that useful? I get the whole linkblog idea - point to interesting content with (maybe) a sentence or two about each one. But a blog made up of some of the inbound emails? It's the commentary from the blogger in question that interests me....

Sci Fi Wire reports that a new Farscape mini-series is in the works. Now, if only it doesn't get as downright stupid as the last season was....

Clay Shirky has a fascinating article up discussing what he calls situated software - software that doesn't attempt to be al things to all people or scale - software that makes use of tacit assumptions in a small group setting:

We've been killing conversations about software with "That won't scale" for so long we've forgotten that scaling problems aren't inherently fatal. The N-squared problem is only a problem if N is large, and in social situations, N is usually not large. A reading group works better with 5 members than 15; a seminar works better with 15 than 25, much less 50, and so on.

This in turn gives software form-fit to a particular group a number of desirable characteristics -- it's cheaper and faster to build, has fewer issues of scalability, and likelier uptake by its target users. It also has several obvious downsides, including less likelihood of use outside its original environment, greater brittleness if it is later called on to handle larger groups, and a potentially shorter lifespan.

I see my students making some of these tradeoffs, though, because the kinds of scarcities the Web School was meant to address -- the expense of adequate hardware, the rarity of programming talent, and the sparse distribution of potential users -- are no longer the constraints they once were

You really have to read the whole thing - the simplest summary seems to be that you don't need to always make software general - in many situations, you can take advantage of the social context the software lives in without trying to generalize it

If the work being done on things like Web Services, CORBA, related technologies is of interest to you, then you'll not want to miss this talk

Cincom Smalltalk Protocol News
presentation
Leonard Lutomski, et al.: Cincom
Tuesday 4:00:00 pm to 5:30:00 pm

Abstract: Cincom Smalltalk protocol developers will present the latest advances in the communication and security frameworks of Cincom Smalltalk. The highlights will include live demonstrations of ObjectStudio-VisualWorks interoperation over Opentalk, Opentalk-IIOP, new Web Services tools, the new ASN.1 marshaling framework, SSL/HTTPS, and more. At the end, the protocol developers will outline and discuss their future plans with the audience.

Bio: Leonard Lutomski manages the Cincom VisualWorks Protocol Team. Martin Kobetic, Tamara Kogan, and Sean Glazier are senior VisualWorks developers and members of the Protocol Team. Martin concentrates on networking frameworks and security; Tamara is responsible for WebServices and NetClients; Sean addresses both security and VisualWave. Andreas Hiltner is a senior Cincom ObjectStudio developer, active in advancing the ObjectStudio Opentalk port.

See you in Seattle!

This interview with Jonathan Schwartz had to be painful for Schwartz, based on some of the comments he (and other management at Sun) have made about MS in the past.

Scoble is impressed by such small thing...

It adds a new folder to Outlook. OK, let's say Bill Gates emailed me right now. Let's say I wanted to post that email out to the world? In the old world I'd need to open up Radio UserLand, copy the email, clean up the HTML, then click post.

New "Kunal's" world? Copy the email over to the folder. It's posted. Done. No more work.

Umm, color me underwhelmed. Sure, it's nifty. But it's hardly that big a deal. It's also specifically tied to a mail client that I wouldn't use if you paid me - Outlook. Not to mention that there are very, very few emails that I would want to post to my blog unchanged - under virtually all circumstances I would want to add comentary of one sort or another. One more thing - if the blog server has a mail interface, one could accomplish the same thing via forwarding - from any email client on any platform. Yawn...

Need to deal with hardware in your Smalltalk application? If so, see what Dan Antion has to say about Bar Code Readers - register today to see what Dan's on about:

SmallBars - a bar code library for Smalltalk
presentation
Dan Antion: American Nuclear Insurers
Tuesday 4:45:00 pm to 5:30:00 pm

Abstract: The bar code library includes support for UPC and Non-UPC linear bar code symbologies. Each symbology is represented in the library as a distinct class, and can be created from strings by a variety of convenient class methods. Every symbology includes support for calculating required and optional check-digits. Symbologies also support the conventional display of human readable characters. The library generates bar codes as device-independent bitmaps so they can be easily printed, displayed or transferred via the Clipboard. The library does not require any installed or licensed fonts. The library was developed for VisualAge Smalltalk and is being ported to Dolphin Smalltalk.

UPC Bar codes include:
    EAN-13	EAN-8
    JAN-13
    UPC-A	UPC-E
    BOOKLAND
Non-UPC bar codes include:
    Codabar	Code 11
    Code 39 (also known as Code 3 of 9)    
    Extended (full ASCII) Code 39
    Code 93
    Code 128	EAN-128
    Standard Code 2 of 5
    Interleaved Code 2 of 5
    MSI (MSI/Plessey)
    POSTNET

The demonstration will discuss bar code technology, the more popular symbologies, library design and usage.

Bio: Daniel Antion is Director, Information Services for American Nuclear Insurers. He is responsible for all systems development activities in addition to managing general information technology efforts. Prior to joining ANI, he worked in systems development for several companies. He also worked as a consultant for Coopers & Lybrand and KPMG Peat Marwick, specializing in information systems and consulting services to financial institutions. An enthusiastic fan of Smalltalk, Dan has made presentations at Smalltalk Solutions and OOPSLA.

See you in Seattle!

Avi wants to have people refer to Seaside (and similar web frameworks as "modal web development" instead of as "continuation based". Has the term "continuation based" already gotten too much buzz to make such a change? I guess we'll find out. I mentioned Seaside a few times at ot2004, and described it as a "continuation based" framework. I guess I helped the old meme along :)

I guess I'm not the only one who was cynical about the Sun/MS agreement - Wired ascribes it to Sun's desperate position in the industry. The Register characterizes it as "Sun throwing in the towel". Just ask yourself - would a Sun that had experienced rising revenues for 11 straight quarters (instead of declining revenues, as they have) - have made this deal? I think not. It's no coincidence that this announcement came on the heels of a 9 percent layoffs anouncement. I'm also somewhat surprised that they promoted this genius to be President and COO. Maybe Schwartz can figure out how to stop Java from being a money spring for anyone but Sun - but I have my doubts about that.

An interesting side note on that point is this from the Wired article:

Daryl Plummer, chief fellow for emerging technologies at the research firm Gartner Inc., said the news that most concerned him Friday was that Sun's 9 percent staff cuts would be across the board - even in research and development, which Sun has fiercely protected in previous rounds of layoffs.

"Cutting R&D is a sign of fiscal responsibility," Plummer said. "However, any cuts in R&D will be unfortunate since Sun relies so heavily on innovation, and they're trying to change the mind-set of the industry - simplifying architecture, getting operational flexibility from chip design to software innovations. It's important they keep that innovation alive."

I've been wondering for years as to when the finance guys would finally blow a gasket at Sun; it looks like that happened. Not necessarily a good sign at this point, and the way Sun did cuts does not, to my mind, usually pan out. Giving everyone a haircut is a great way to cripple the good with the bad. Far better to pick winners and losers, and just shoot the losers. The more I watch Sun, the more I see a bigger version of ParcPlace Systems - losing money hand over fist, and utterly baffled as to what business they are in...

I really, really wish that people would stop using the term loose typing when they mean dynamic typing. ISerializable has a post mentioning Python that way this morning....

Come hear an experience report on deploying Smalltalk apps in an enterprise system at StS 2004 - Register now:

Pragmatic Enterprise Software Delivery
presentation
Angus MacArthur and Sean Morrison: Ontario Teachers' Pension Plan
Tuesday 4:00:00 pm to 4:45:00 pm

Abstract: Ontario Teachers' Pension Plan (OTPP) is the pension fund for all of the public school teachers in Ontario, Canada. The Plan services 250,000 current and retired teachers and controls 68 billion dollars (CND) in assets. The internal software development team builds and supports multiple applications, most of which are variants upon a common base. These applications serve hundreds of internal and thousands of external users.

The presenters intend to introduce a sampling of the tools and processes they have developed (in Smalltalk) that alleviate the tedium and risk in the building, testing, and deploying of software. Specific topics will include software QA, automatic build creation and configuration management.

Bio: Angus MacArthur has been working in Smalltalk for over a decade in a variety of industries.

Sean Morrison has been developing software in Smalltalk and Jave at OTPP since graduating from the University of Ottawa in 1999. He is also active in the Toronto Smalltalk Users' Group.

Scoble points to the much publicized Sun/MS deal - and blogosphere reaction is very positive so far. My read? This is very much like the MS/Apple deal - it will have the effect of propping up one of MS' competitors enough to keep the justice department at bay. One things for sure - MS has been truly blessed in terms of who they have had to compete with :)

Update: This post from Ted Neward is very representative of the general comment trend. The basic thought seems to be "they made up and are playing nice". Nope. Let me boil it down

• Sun lost (count the declining revenue quarters and the headcount cuts)
• MS won - and to prevent the Justice department from seeing it as killing a competitor, they are paying to prop Sun up. You know, just like they paid Apple to prop them up

This is easily played as a win-win, but what it amounts to is the best that Sun could get from a really, really bad situation

Mark Pilgrim posts on the trend towards backchannel conversations during a presentation (typically IRC or IM). This is nothing new; I have IRC and IM conversations going on all the time during phone conferences. To a large extent, this is the responsibility of the speaker(s). If they can't keep the audience's interest, who's fault is that? Certainly not the audiences. And what about blogging a session? I've just blogged ot2004, and that's certainly a form of back channel conversation. This point stands out to me:

I can not be any clearer: I wholeheartedly support this. Despite hysterical objections from the usual suspects, I have seen the benefits of the backchannel firsthand. At ApacheCon last fall, Ken Coar announced during the initial keynote that there were IRC channels set up for the conference (one for each presentation room, and a main one for the conference in general). When I presented, I went so far as to put the address of the IRC channel on my first slide, to remind people where they could talk about me behind my back for the next 45 minutes. A friend in the audience forwarded me a copy of the channel transcript afterwards, and I discovered that several of the best questions came out of discussion in the backchannel.

There are people who know, and there are people who raise their hand and ask questions out loud, and they aren't necessary the same people. But if these people get together in real time, a question gets asked that otherwise wouldn't get asked, and it gets answered, and everyone benefits. Real-time is important; after the presentation ends and everyone shuffles off to the next one, it's too late.

My experience with public speaking is similar - there are people who won't ask questions in public, and people who will. A back channel provides one more way to surface some of those hidden questions. Anything that creates more conversation can't be all bad :)

Despite all the revelations about old AT&T , Novell (etc) Unix related stuff, SCO still thinks they can win. Now, IANAL, and I've seen judges and juries do some, ummm, interesting things. Read the interview.....

PR Opinions points to the interest and influence (still unclear) of blogs in the tech journal sector. Personally, I find that I'm reading fewer journals...

"We literally have a daily discussion on this topic. We have a lot of top people writing columns and perspectives but we know that the bulk of our readership still comes to us for the news. In one sense, my editor for opinions, Charles Cooper, writes a column that's sort of a blog, and we have this "Your Take" feature now, whereby readers can respond to his column, but it is not a two-way thing. I think with news the question is which blogger do you really trust. There are so many of them. It seems to me like it's a pretty incestuous thing going on. One journalist points to another journalist's blog and other bloggers point to other people's blogs, and you somehow think that this is the most popular thing. But is it credible? I don't know."

Interesting to see how this plays out over time.

Well, I thought this was funny :)

So I get home, greet the family, boot the laptop. It refuses to find the WiFi in the house. Check the other system that uses WiFi - yes, it sees the network. Pull the card in and out of the machine a few times, swear a lot, swap cards. Nothing. Disable the wireless card, enable it, pull it out, put it back in, nothing. Then, after 45 minutes, for no apparent reason, it finally spots my network. The maddening part? It was offering to connect to my neighbor's network the entire time. Grrrrrr.....

Well, almost back home. I'm tapping this out in a cab while it crawls through rush hour traffic on the BW parkway - I arrived at DCA not too long ago. It's a rainy, nasty day; it was actually nicer in the UK (London) this morning when I woke up. That was 12:30 am local time, so I've been up a long while now as well - adding to the generally odd feel that these long travel days acquire. No luck with the wardiving while I've been stuck in traffic either - I've stumbled on one WiFi spot, and that was WEP protected. So much for moblogging from the cab :)

One thing's for sure; this ride back home shows me why I'm so glad I work out of a home office most of the time - I can't believe that people put up with this kind of traffic on a daily basis!

Spotted this on edit and continue in jaybaz_MS's WebLog:

You now have a situation where a simple unit test is only a couple function calls away from any bug.  Compare to the Einstein case I mentioned earlier, and you'll see that complex repro scenarios are comparitively rare.  You can reproduce the faliure in seconds - just run the unit tests.

When that cycle is so short, E&C doesn't help nearly as much. 

The "Einstein" reference is in relation to an internal to MS view on various developer types; I'll have to post on that at some point. Here though, I see something interesting - MS is only just starting to support the kind of immediacy (edit and continue) that Smalltalk supports so much better than they do - and their developers just don't see the value. Amazing. I used a TDD approach when developing my MarkupHelper (for the blog posting tool), and it was so much easier to do since I could look at stuff in the debugger and change it right there - without having to run back to some different tool. Interesting difference in viewpoints at the very least...