PLEASE CLICK ON AN IMAGE TO ENARGE I got in a little late. Keven Appert let me in the door, because the door to Cogswell College is locked on weekends. He showed me to the meeting room. Kevin happens to be facilitating the meeting. There were almost no seats left available, and I had to sit way in the back. There was about 50 people in attendance at Forth Day altogether. As I entered the meeting room, Brad Nelson was giving a demonstration of Rainbow Forth, a Forth written in JavaScript. Apparently with this it is possible to use Forth in place of JavaScript, although I would assume that there would be a speed penalty. After this we had a meeting break during which I introduced myself to Chuck Moore and Jeff Fox. Jeff said that he had a fantasy about using an entire wafer of Forth chips as a single computer, then using a dozen of those in a tower. If there are 1000 Forth chips on a single wafer, each chip contains 40 cores, and there are 12 of those in a tower, that equates to 3 quadrillion Forth operations per second. I have no idea how this could actually be hooked up. Jeff mentions that it is a big step from prototyping to fab, and that they have more experience in this area now. I get Chuck's autograph. I've wanted to meet him for years. Then the meeting starts again and C. H. Ting gives a talk about eForth. eForth is a system which is easily ported to just about any hardware. For example there is a version for the ARM7 on Nintendo Gameboy Advance. Ting described his different uses of eForth for different hardware. After this Leon Wagner, the President of Forth Inc. speaks. Leon develops Forth applications. First he talks about the Forth Standard. There was a Forth Standards meeting in Europe. He says that the European Forth community is very vibrant. He also says that the German group is very vigorous. Leon also talks about Forth Inc's Linux Forth. Venture Forth for the C18 works on both Windows and Linux. The Linux Forth is still in Alpha, although it is stable. Forth Inc. is using their Linux Forth for a project using Forth chips in conjunction with a Linux server. Next up was Samuel A. Falvo, creator of the Kestrel computer. He was experimenting with the Forth chip and video. He says that the 24 core chip can be used for IBM VGA at 640x500 resolution and that a 40 core chip could be used for SVGA. He mentions that what he is doing is similar to what Atari and Commodore did in regards to graphics. He was using several cores to generate VGA sprites, while using NO external memory. In other words, he was using computation alone to generate sprites, rather than using a frame buffer. Insert the sound clip from the videogame Gauntlet, "That Was A Heroic Effort" This was because he only had a thumb drive to work with, instead of a 40 core devkit. On the subject of a Forth chip based graphics board on the order of something from NVIDIA, it was mentioned that it can be done, but that NVIDIA is more optimized for this sort of thing. Maybe a 9x9 matrix of Forth chips could do it. Samuel does mention that the C18 is an incredibly stable system. Afterwards Joe Connor spoke. The subject is Creole Forth. On his first attempt he went about it the wrong way and used a Delphi database for the dictionary. But this has now been rectified. Next up was Dylan Smeder, marketing director of Intellasys. He says that last year they showed the 24 core chip. Now they have the 40 core chip. It has an extra analog core. There is a redesigned RAM server as well. There is now a new development kit for the 40 core chips. It includes a serial module, an audio module / midi board, a USB port, and a level shifter. What is in store for the future? There is a new device on the way with more than 40 cores. Also, Venture Forth 2.0 is coming. This will include automatic documents downloads and automatic upgrades. Next to speak is Charley Shattuck from Intellasys. He demoed Venture Forth operating in Linux. He says that he no longer likes monochrome listings. He also mentions that the alpha version of Forth Inc's SwiftForth works well enough to not be alpha. The proceeding speaker is John Rible. He discusses eForth on the C18. One of the features of eForth on the C18 is that it can address up to 32 megabytes of RAM through bankswitching. eForth has 256K. This is split into two 128K areas. One of those areas is paged across 32 Megs. After this we listened to Jeff Fox. One topic was the issue of lockup. Lockup is a big issue in the field of parallel programming. His solution to this is to teach people how to write software. Towards this Jeff is now using datagram comments for each parallel processing word, which is similar to a stack comment, except that while stack comments are used between words in a single microprocessor, a datagram is used between words resting on separate microprocessors in a matrix of units. As long as data flow diagrams are complementary lock ups can't happen! Then Jeff talked about an image processing application for collision detection for an automobile. This application must watch out so that you can't hit somebody while you are driving. Next in line is Michael Monvelishky. He showed his modular designed software for his music application. Afterwards was Dave Jaffy who gave us all a handout pertaining to a who's who project for FIG. You send them information about yourself and your interest in Forth and they compile a list of who's who on the FIG website. FIRESIDE CHAT Chuck Moore now speaks. This is the 40th Anniversary of Forth. Chuck spoke about the history of Forth, when the first Forth was created, and such. Forth started life on an IBM 1130 minicomputer. It was from the Burrows 5500 mainframe that Chuck gained Forth's stack archetecture. He mentions Mohasco, and the work with radio telescopes at National Radio Astronomy Observatory (NRAO). Then there was Forth Inc. After awhile he became fed up with hardware because there are many badly designed circuit boards. Then there was the progression of Forth chips. Now we have 40 core chips. Chuck talks a bit about the C18. He says that one must do shuttle runs before full production of a microprocessor, because you can't always trust the fab to do things right. There are 10,000 40 core chips in existance as of the time of the meeting. The remainder of 75,000 chips will be comming very soon. All of the bugs of the C18 core should now be worked out, thus the chip should be perfect. Chuck wants to create a little board which will have a 3x3 array of 40 cores on it. He might put OKAD on it. There are many, many multicore chips on the way from different companies now. They are all more expensive than the Intellasys products, and they all require a heat sink. Intellasys leads the way. There exists the potential of a 3 to 4 times speed up with the latest fab process and transistor size, but it isn't necessary at the time being. If it is requested it can be done. Right now the C18 runs at 700mhz. It was a nightmare to come up with the right package for the chip. The industry is in severe recession. If this chip is successful, then there will be a resurgence of popularity for Forth. Chuck then discusses the instruction set of the C18. He says that it is a fun instruction set. He walks us through each of the instructions on the computer. Why an 18bit microprocessor? In the old days they had 18bit computers and 18 bit RAMS. Also, 16 bits is a little too tight. After the meeting, many of the Forthers go out to a Chineese restaurant and it was yummy! Here I am pictured with Kevin Appert And by the way, the Forth chips are not vapor. Here is a picture of a 40 core devkit.