Oct 24th 2001.	Camp Smalltalk Essen ESUG 2001

Back to Day 1.

It's morning, the second day of Camp, which started before I did this morning, but that is ok, none of this OOPSLA 08:00 hour start times please! Arriving at mid-morning I find only a few free seats within the two large rooms, this is great lots of Smalltalkers at work, how often can you see 30 plus people do pair programming in Smalltalk?

I attempt to get an Internet access to ensure my posting last night actually worked. This is a bad decision, after much fiddling we found that the win2k box we had only handed out 24 IP addresses, then you need to buy the win2k enterprise server edition. This was a very frustrating thing to deal with and to figure out, only Microsoft could get away with fostering such a thing off on the computer community. The macintosh users were late arrivals, mmm artistic people need more sleep after a night of drinking, so it first appeared that only macs weren't able to connect. This issue was solved by handing out static IP addresses and hoping for the best. Perhaps the DHCP server would check for collisions before give out what would be a duplicate address. A note to past/present Camp Smalltalk organizers, perhaps a small green book on lessons learned is in order for future Camp Smalltalks.

A quick check around the room showed that Stephane was working on getting an image together that showed various interesting Squeak projects. At the moment he was looking at how the "jump to project" logic worked. I mentioned the work Henrik Gedenryd had posted on " [Modules] From here to there" late last night so we pulled the change sets from the Squeak list and started exploring. This took about half an hour before we realized that it would not solve the problem Stephen was hoping to solve.

FROST 3 plus 4 works was the yell from the other side of the room, and more on that later.

Did I mention the morning started late? So lunch is where all thirty plus of us endured poor service at an Italian German restaurant, however I had some interesting discussions with Joseph about tethered Smalltalk images. His vision is to have a very small basic image that supports only the kernel, and the socket layer. By talking to a fully functional image that exists on the Internet somewhere it could request that server to load modules from a repository, compile them and sent it the bytecodes to install. That way the small image could start from a very small size, and add functionality to meet it's business needs, and forgo loading dangerous things like compilers.

After lunch it was apparent that frantic effort was required, Camp is half over, really I think Camp should be a week long, so most campers put their heads down and keyboards churned, in fact we were making progress. Later the suggestion was made that some people would be willing to attend the free camp during the Conference and forgo the conference because of the expense, that would extend their experience. Would that work? Others can ponder the decision.

GLORP, still had a large problem exchanging source files. But the Dolphin port is coming along, we are making progress, we'll even have an SAP interface. The GLORP group told me they did a fileout from VisualWorks then VisualAge and into Dolphin manually that worked ok, in fact Dolphin didn't have any problems. 138 out of 153 tests past then Dolphin "screams into the weeds". Dolphin doesn't have a method isString, you must be kidding was the reply, tedious but steady effort. An hour later the Dolphin folks told me they have finished the port, so in fact they were pleased with their progress.

Web Services, who's got the SOAP.

Over in the Web Services group Mark was encapsulating SOAP, at this point in Camp Smalltalk VisualWorks has hit VisualAge with a web services request, and soon this will have the roles reversed. Then work will be done on Squeak. One of the interesting things I thought was watching various firewall software complain about incoming port requests.

The Refractory Browser porting group was hard at work, the first one or two windows have been done, some discussion about test cases for the GUI was made, but it became apparent this was very difficult to do in Morphic and this effort was postponed. The first window does rename of methods, now they are working on the GUI for extract method. The major problem has being getting going, also lots of time wasted considering GUI Morphic test cases, which proved to be too difficult. But "we'll be here the whole weekend" because this is an important task. Claus at the same time was working with the Refractory Browser test cases in Smalltalk/X to ensure they worked correctly. Over in the corner another few Smalltalkers were working on the rename instance variable window to work, yes someday soon we might get a version for Squeak.

Claus of course was distracted in his work and showed us a oscillator he created, which he tied to a oscilloscope and drove from an frequency oscillator. Ha, very cool. Think of various stereo devices visually displayed with analog knobs and meter feedback with the ability to connect virtual wires between them, then observer the visual results and the sound! I did ask him which client wanted this, but the answer was no-one but it was neat to do, so I suggested he port the Mpeg plugin to Smalltalk/X. Interested parties should email Claus if they think this is a good idea. Some of the Squeak Morphic guys should think of tube stereo devices and Morphic and the interesting things they might visually do.

The NUMERIC group had reached chapter 7. Authors of books should always give examples, because when readers want to try the code, then er, it should be simple right. Ah, but maybe there are some details left up to the reader to resolve, and yes, the examples should at least compile. The group tells me it takes 5 minutes to file in a section then three hours to fix problems with the code. Slow but it's not too difficult. Over the day after various smalltalkers around the world read the Camp Smalltalk report last night this group started to receive change sets and pointers to information about SUnits. I was told in the evening that work originally started in 5i.3 but a decision was made to move to 3.x because of name space issues.

Carsten was working on the FROST version, the first trial was getting a version that was close to 5.i A partial version came from Florin Mateoc, and Kerry . In the morning as mentioned they got 3 + 4 to work. One of the odd things was the class names had slash in them. They had to make a change to ENVY to import the classes with slashes, further work will be required to rename the class names to remove the slashes. Right now they were setting up test cases in SUnits to run. More complex examples have parsing errors due to changes between 3.x and 5.x with the advanced tools compiler. Slow but sure progress, but they couldn't tell me how fast 3+4 actually ran.

Stephane was crashing the macintosh VM. Interesting to me, but not really a Camp Smalltalk activity, I hope. This was part of the activity he was doing in putting together a set of projects he could have on disk that he could load into an image to show what Squeak could do. I think he is hoping to put this on the ESUG CD. As this was going on the ESUG organizers were abuzz with the numbers of confirmed attendees, they were much higher than expected into the 3 digits range. Wow, in fact a new record for ESUG.

Craig Latta was here and was working with the standard socket group by integrating Flow into the work they were doing. Also for the work Joseph was doing with modular squeak in stable smalltalk Craig was looking at assembling a change set with the primitives and the ExternalStreams interface.

The Smalltalk isolated dialect group was retrieving the work Paul Baumann (pbaumann@effectnet.com) had done with SRP http://www.effectnet.com/pbaumann/index.html. They had been working on suggestions by Peter for the ANSI conversion stuff, obviously this was just a subset of work required for dialect conversion. Still more work to determine what needs to be in the LCD Smalltalk and what is in ANSI stuff. Much of the effort was still resolving what the solution should be and collection information from Smalltalkers around the global on possible solutions.

The web services group by this time was hard at work at a blackboard working issues with their framework and planing what next to tackle, and beside them Marica and Peter were working on something to replace blocks, or be an alternate to blocks. They were working a system to implement the specification of the Z mathematical notation language. Although Smalltalk has blocks, it wasn't powerful enough so they were extending things.

By this time Craig and Mark of IBM had finally gotten together and connected. Mark had realized that Craig was "The Flow Guy", something he was interested in, perhaps for the web services project?

Meanwhile Stephane was working on his image, he had determined that there was a problem with the flash player and if you attempted to duplicate a flash object the image would become unresponsive, yet you could still do a file quit. I felt this was a problem with memory so we resized the VM memory size to see if this would fix the problem, this did (usually), Squeak doesn't handle low memory conditions very well, something that needs looking at. However Stephane could now show this image Wed morning.

Craig and Joseph were still working on porting Flow to Stable Squeak. Although this had been mostly done via work by John Sarkela, myself, and Craig there still was integration to be completed using the latest version of Craig's work. Right now they were cleaning up a file out, and at which point it was clean they would start SUnit testing, I pointed out that I had done an SUnit test suite when I had ported Flow to the SqC socket primitives. But at this point they were struggling with removal of some instance variables that now were obsolete, but in their version of emac the swiss keyboard mapping was making things difficult.

19:30 hours so soon? Time to find a restaurant and again 25 of us found a German Mexican restaurant and tax their data processing capabilities to death. Over some lightly spicy Mexican food I found out that Joseph Pelrine wants to coordinate the writing of a book on Smalltalk VM experiences learned over the last 30 years plus. We all thought this was an excellent idea. If you would like to give input then please contact Joseph directly, if you can twist the arm of some retired smalltalker, then again talk to Joseph. As posting to various lists alerted us, yes time is running out, some people do need to document their thoughts.

Did you know how Smalltalk makes steel? Or how it affects the painting of your Honda car, if you attended Camp Smalltalk you would know. I must say over dinner we managed to drag a few words out of Joseph about his work on modular Smalltalk within Stable Squeak, but perhaps not enough to spoil his upcoming presentation to us all. Basically he has a vision about how images can be made, drawn from Allen Wirfs-Brock work, and his experiences in delivering mission critical smalltalk systems. I'll have more later in the week, but it sounds like much of the hard work has been done and just further refinement is required to add things like the repository backend. Mind GLORP would fill this void quite nicely.

As the evening ended, some Americans cast off to find an ATM. A difficult task in this city on a Sunday. Other found an American sports bar by Kennedy Square, and some of us retired to do some more typing.

Forward to Day 3.