Volume 7, Number 52 24 December 1990 +---------------------------------------------------------------+ | _ | | / \ | | /|oo \ | | - FidoNews - (_| /_) | | _`@/_ \ _ | | FidoNet (r) | | \ \\ | | International BBS Network | (*) | \ )) | | Newsletter ______ |__U__| / \// | | / FIDO \ _//|| _\ / | | (________) (_/(_|(____/ | | (jm) | +---------------------------------------------------------------+ Editor in Chief: Vince Perriello Editors Emeritii: Thom Henderson, Dale Lovell Chief Procrastinator Emeritus: Tom Jennings Copyright 1990, Fido Software. All rights reserved. Duplication and/or distribution permitted for noncommercial purposes only. For use in other circumstances, please contact Fido Software. FidoNews is published weekly by and for the Members of the FidoNet (r) International Amateur Electronic Mail System. It is a compilation of individual articles contributed by their authors or authorized agents of the authors. The contribution of articles to this compilation does not diminish the rights of the authors. You are encouraged to submit articles for publication in FidoNews. Article submission standards are contained in the file ARTSPEC.DOC, available from node 1:1/1. 1:1/1 is a Continuous Mail system, available for network mail 24 hours a day. Fido and FidoNet are registered trademarks of Tom Jennings of Fido Software, Box 77731, San Francisco CA 94107, USA and are used with permission. Opinions expressed in FidoNews articles are those of the authors and are not necessarily those of the Editor or of Fido Software. Most articles are unsolicited. Our policy is to publish every responsible submission received. Table of Contents 1. ARTICLES ................................................. 1 Math Co-processor Options ................................ 1 Aviculturists Wanted! .................................... 9 Some More about the 'GhostNET' Echo ...................... 10 PUBNET -- New Organization for Computer Publishing ....... 11 The Respectable SysOps Guide ............................. 12 NewStyle Packets ......................................... 15 2. COLUMNS .................................................. 25 The Captain's Log: Stardate Christmas 1990 ............... 25 3. LATEST VERSIONS .......................................... 27 Latest Software Versions ................................. 27 And more! FidoNews 7-52 Page 1 24 Dec 1990 ================================================================= ARTICLES ================================================================= MATH CO-PROCESSOR OPTIONS by Dave Appel, USPA license # C-20451 Just a dumb ascii user on 1:231/30 C:\INTRO\MATH Here is a list of "things you should know" about upgrading your IBM or clone computer. I am going to focus on upgrading with a numeric co-processor, a math chip. And, whether or not you really need to upgrade to the latest 80486 system even if you think you are maxed out on your 33 Mhz 80386. Even if you are familiar with the 80x87 series read on. Now there are alternative clone chips. One of these alternatives may just save you the expense of upgrading your whole motherboard or system. The thing that got me interested in this topic was a July 1990 Byte article reviewing two 25 Mhz 486 systems. The big news is that a 25 Mhz 486 CPU is no faster than a 33 Mhz 386 CPU. The only benefit that you get for spending thousands of extra dollars for the 486 system is that the 486's built-in floating point unit (FPU) is about twice as fast as the 33 Mhz 80387 (that you can add to your 386 system for about $550 to $600.) If you are maxed out with your current 80x86/80x87 combination you have several alternatives besides upgrading to a faster motherboard or even the 486 system. Now that I've told you what my thesis is, let me back up a bit and explain the basics. C:\BASIC\MATH A "numeric co-processor" or "math chip" is an integrated circuit chip that you can plug into your motherboard. Its sole purpose is to speed up numeric calculations. This chip does calculations which would otherwise be done with software running on the CPU chip itself. Numeric co-processors only benefit a small number of users because there are two conditions that must be met in order for these chips to make a difference. One is that a good portion of your computer time is spent doing mathematical calculations. The second is that the software doing these calculations must be smart enough to detect the presence of a co-processor and use it. Computer Aided Design (CAD) programs such as AutoCad usually require a co-processor in order to get anywhere near acceptable response time. Lotus 1-2-3, QuickBasic, PageMaker and Ventura are other common programs that can take advantage of the co-processor. (Ventura and PageMaker only take advantage of it when printing graphics.) FidoNews 7-52 Page 2 24 Dec 1990 If you are running one of these programs, and you can ascertain that the program is "compute-bound" as opposed to "I/O-bound" (such determination is a whole science unto itself) then ask yourself if the computations are "number crunching" or "character crunching." If your application is compute-bound while doing mostly numeric calculations, then adding a numeric co-processor or upgrading to a faster alternative co-processor will allow your programs to run much faster. If you are compute- bound, but the application is just moving data around and not doing actual numeric calculations then a co-processor will NOT make it run faster. One application where a co-processor is useless is a dedicated file-server. Although the 80486 CPU is now touted as "ideal" for file-servers, it is a waste of money for such an application. The extra money you spend for the 80486 is better spent on faster disk drives or more memory for buffers. File- servers don't use the floating point unit. The claim to fame of the Intel 80486 CPU chip is that it has the numeric co-processor built-in. You don't need to add one. If you have an 8088, or an 80286 or an 80386, there is probably an empty chip socket on your computer's motherboard where you could plug in the co-processor. The part number of the co-processor depends on the part number of your CPU. The price depends on the speed (Mhz rating) and the chip maker. CPU NUMERIC PRICE RANGE(depends on speed) CO-PROCESSOR 8088 8087 $ 80-$150 80286 80287 $170-$250 (AMD's=$99) 80386 80387 $280-$600 80386SX 80387SX $280-$400 80486 built into CPU chip In addition to matching the part number, you also have to match the speed. If you have a 10 Mhz 80286, you need a 10 Mhz 80827. If you have a 12 Mhz 80286, you need a 12 Mhz 80287. If you have a 16 Mhz 80386 you need a 16 Mhz 80387, etc. (Assuming you "need" one in the first place). If you have a 10 Mhz 80286 CPU and a 10 Mhz 80287 co-processor just swapping out one of these 10 Mhz chips and replacing it with a 12 Mhz chip won't do you any good. The timing controls are in the motherboard itself. Adjusting the timing controls won't do any good either, because the other chips on the motherboard are matched to the timing too. (This is akin to those who think that swapping their 120 ns memory chips with 80 ns chips will speed up their system. It just doesn't work that way.) C:\NEW\MATH There are five options for those who would benefit from adding a co-processor (compute-bound doing numeric calculations). They are: FidoNews 7-52 Page 3 24 Dec 1990 Vendor Parts ------ ----- Intel 8087, 80287(-XL(T)), 80387, 80387SX IIT 80287, 80387, and 80387SX replacements Cyrix 80387 and 80387SX replacements AMD 80287 replacement Weitek non-80x87-compatible co-processors for 386 & 486 Intel, IIT, AMD, and Cyrix co-processors can plug right into the socket on your motherboard. Just be sure to choose the part number and speed that corresponds to your CPU. The Cyrix and IIT chips are about 2 to 3 times faster than their corresponding Intel part. Intel recently replaced their original 80827 chips with the newer 80287-XL and 80287-XLT. They claim the 80287-XL is 50% faster than the original 80287 of the same Mhz rating. But the IIT chip is still faster. The 80287-XL will work with any 80286 CPU up to 12 Mhz. According to Byte benchmarks, the AMD chip is slower than the Intel chip but only 1/2 the cost. The AMD chip is plastic, wheareas the other makers use ceramic chips. The Cyrix and IIT co-processors are also slightly less expensive depending on where you buy them. In addition to being "pin compatible," the Cyrix, AMD, and IIT chips are "code compatible" with the Intel chip. Your hardware and software cannot tell them apart. Just plug them in and go. It's the same as the clone philosophy of marketing. You have to be better and cheaper than the "big guy" in order to get market share. Interesting side note: Intel didn't start agressively advertising their numeric co-processors until IIT and Cyrix came on the scene. Amazing what a little competition will do. If you already have an Intel 80x87 chip, then switching to either the IIT or the Cyrix equivalent might get you the extra number crunching power you need. This could save you the cost of upgrading to a faster motherboard or a whole new system. There are several "gotchas" here. Gotcha #1: Intel somehow fixed or boosted their 33 Mhz version of the 80387. While the Cyrix chip is at least twice as fast as the Intel chip in the 16, 20, and 25 Mhz versions, the 33 Mhz version of the Cyrix is only marginally better than the 33 Mhz version of the Intel chip. But... the 33 Mhz Cyrix is still 2.5 times as fast as the 33 Mhz Intel 80387 when doing transcendental functions such as sine, cosine, etc. This makes it a rough equivalent of the 80486 at this specific type of calculation. Gotcha #2: The Cyrix chip has a 90 bit internal architecture whereas the Intel chip has an 80 bit internal architecture. This means that on certain intensive double- precision calculations, the Cyrix chip is SLIGHTLY more precise, or better, than the Intel chip. Cross-machine consistency may be more important to you than unnecessary extra precision. It is possible that after thousands of iterations, the Cyrix will give a result that is more precise than the Intel's result. Both chips are IEEE compatible and give "accurate" results, but "more FidoNews 7-52 Page 4 24 Dec 1990 precise" is still different. Sidenote: This is the issue on which Intel charged that the Cyrix chip is "not Intel compatible." This was a cheap shot since the Cyrix chip is better in this regard. But... very few applications require the precision where the difference would be notable. Single-precision and most double-precision calculations would not be affected at all. So, if cross-machine conformity is an issue, and you already have one vendor's co-processor installed on some machines, you might want to stick with that vendor on others. If you only have one machine, or haven't installed any co- processors on your machines yet, or are going to upgrade all of them at once, the Cyrix or IIT is still more bang for your buck. Gotcha #3: This applies to all co-processors including Intel. When motherboard manufacturers test their boards prior to shipping, almost all of them neglect to test the co-processor socket. This is probably because no motherboards are shipped from the manufacturer with a co-processor already installed. This results in some being shipped with defective sockets or socket wiring. You might insert your co-processor chip and think it is the fault of the chip if it doesn't work. It could be the fault of the socket, its wiring, or the motherboard. One vendor, Microway, offers a diagnostic program to help you pinpoint the problem. Their address is at the end of the article. Gotcha #4: Intel did not ship the 80387 co-processor at the same time as the first 80386 CPUs. Therefore, manufacturers equipped their first 80386 motherboards with an 80287 socket instead of, or in addition to, the 80387 socket. The 80287 is NOT pin compatible with the 80387. They require different size sockets. If you are going to add a co-processor to a 386 purchased in 1988 or earlier, check the motherboard specifications before making a purchase. It might even be a good idea to open the box and inspect the motherboard to visually verify what kind of co- processor socket is there. An 80287 socket is rectangular and has holes for 40 pins in 2 rows of 20. An 80387 socket is square with round holes for 68 pins. It has 2 rows of holes on all four sides. A Weitek 3167 socket is square with round holes for 121 pins. It has 3 rows of holes on all four sides, with one extra hole on an inside corner. The co-processor socket is usually located adjacent to the CPU chip. Gotcha #5: Some of the very latest 286 laptop computers don't use the 40-pin socket for the 80287. Some, such as the Compaq 286/LTE and the Tandy 2800, use a square socket called a PLCC. Intel's 80287-XLT is the only math chip currently made for these sockets. Gotcha #6: Some motherboards drive the co-processor at a different clock rate than the CPU. It will be either the same clock rate as the CPU, or a certain fraction of the CPU clock. Some motherboards have jumper settings to determine the co- processor clock rate. My ATTAK brand motherboard drives the CPU at 10 Mhz. But I can jumper it to drive the co-processor at either 10 or 6.66 Mhz. FidoNews 7-52 Page 5 24 Dec 1990 If your motherboard uses one of the 16 or 20 Mhz 286 chips from Harris or AMD, be sure to check the speed at which the co- processor is driven, and if it is jumper-selectable. For instance, JADO brand motherboards can drive the CPU at 20 Mhz while driving the co-processor at jumper-selectable rates from 8 to 20 Mhz. IIT is the only manufacturer I could find that makes a 20 Mhz 80287 compatible chip. C:\ADVANCED\MATH The Weitek co-processor is a unique and popular chip. It has a dual 32 bit and 64 bit internal architecture. This makes it much faster than the other co-processors. It is slightly less accurate at double-precision than either the Intel or Cyrix. It is still IEEE compatible, but just "different." Again, only the pickiest of scientific applications need precision beyond what the Weitek provides. The Weitek chip is about twice the cost of the Intel chip at each Mhz rating. But, it is even faster than the Cyrix and IIT chips. It is 3 to 5 times faster than the 80387 of the same Mhz rating. And here is the kicker, the 33 Mhz Weitek 3167, coupled to the 80386, is FASTER than the 80486's internal FPU on everything except double-precision matrix (vector) calculations. The manufacturer claims that a Weitek 3167 coupled to a 80386 provides the floating point speed of DEC VAX 8650. If the internal floating point unit (FPU) on the 80486, or the 80386/Weitek combination is still not enough power for you, you could add the Weitek 4167 chip to a 486 system. The Weitek 4167, coupled with the 80486, more than doubles the floating point speed of the 80486 alone. (The Whetstone rating for the Weitek 4167 equals the speed of many of today's mainframes.) The Weitek chips have their own gotchas. Gotcha #1: The Weitek chips are NOT "pin compatible" with the Intel co-processor chips. Most motherboard manufacturers are now putting "Weitek sockets" on their motherboards in addition to 80387 co-processor sockets. Acer, ALR, AST, AT&T, Compaq, CompuAdd, Convergent, Dell, Everex, HP, NCR, Packard Bell, Sun, Tandy, Wyse, and Zenith are some that provide this feature. If you are buying a new computer system for number crunching, this may be a feature you should consider. Gotcha #2: The Weitek chip is NOT "code compatible" with the Intel 80x87 chips. In order to get the screaming speed into this chip Weitek had to change both its electrical interface (hence the need for a separate socket) and the software interface. Like all the other "gotchas," this one also has a "but." A growing number of software packages, that are popular among users needing number crunching power, have Weitek compatibility. CADKEY 3, Vericut, ANVIL-5000pc, CADMAX, VersaCAD/386, AT&T Unix, 386/ix, SCO Unix, Phar Lap Virtual Memory Manager, Eclipse Dos Extender OS/386, are some that are Weitek Compatible. Microway also makes Fortran, C, and Pascal compilers, graphics applications, and Math/Stat applications that support the Weitek chip. FidoNews 7-52 Page 6 24 Dec 1990 C:\OTHER\STUFF If you are going to do a lot of number crunching, your system has to get those numbers to the co-processor before they can be crunched. Many factors go into the overall system speed besides the Mhz rating of the CPU. The major factors besides the CPU are the disk and main memory. ESDI and SCSI disks are faster than MFM and RLL. Controllers that support 1:1 interleaving (full track buffering) are faster than controllers that are 2:1. Look for these features when building a system for speed. Memory on the motherboard is accessed a lot faster than memory on an expansion card. Look for a motherboard that has the capacity for all the memory that you expect to need. The amount of memory supported directly on the motherboard varies a great deal from one manufacturer to another. Memory on an expansion board can take 2 to 4 extra cycles to read. "Cache memory," "0 wait state memory," or "system cache" is another feature to look for. This is not to be confused with disk caching. Disk caching is performed by software, and buffers data between the disk and main memory. "System cache" or "0 wait state cache" is a special kind of memory that buffers code and data between main memory (DRAM) and the CPU. This kind of cache is implemented in hardware right on the motherboard. It has nothing to do with disk caching or the disk controller. CPUs are so much faster than dynamic ram (DRAM) memory that "wait states" are used. This has the effect of throttling down the CPU. Benchmarks published elsewhere have amply demonstrated that systems with cache memory are considerably faster than systems without cache memory when everthing else is equal. System cache overcomes the problem of slow DRAM slowing down the CPU. 80386 systems rated at 20 Mhz and up usually come with some kind of cache. Common denominations are 32K, 64K, and 128K. The more main memory you have, the larger you want your cache. To build a faster system, go with the largest cache option. CPUs rated at 25 Mhz and up really NEED this cache or else the speed of the CPU goes to waste. How much cache is needed is a subject of debate. It depends on the application, and the size of the portions of code that get executed most often. Multi-tasking or multi-user systems would need more than a non-multi-tasking system. Heavy CAD graphics also benefit by having cache. Constant manipulation of large matrices in memory, such as large spreadsheets, also benefit by having cache. The 80486 CPU chip has a built-in cache controller and 8K of internal cache. This is one of the reasons that a 25 Mhz 486 can perform at the same speed as a 33 Mhz 386. However, 8K of cache is no where near enough for a multi-tasking system that has megabytes of main memory. So look for a system cache of 64K and up when shopping for 486 systems too. Don't be fooled by Intel's claims of "built-in cache" on the 486 chip. It's nice that it's there, but it's not enough. FidoNews 7-52 Page 7 24 Dec 1990 Here is a table of benchmarks of some of these co- processors. A 33 Mhz Compaq 386 and a 25 Mhz 486/4167 were used: (From the paper "The State of PC Numerics in 1990" by Stephen Fried, Vice president, R&D, Microway, Inc. By permission.) Benchmark Intel Cyrix Weitek Intel Weitek 80387 83D87 3176 80486 4167 scalar S 1611 1613 4050 3307 7635 D 1435 1543 3570 2942 6773 matrix S 866 875 2563 1877 3200 D 672 726 1127 1703 3074 trans S 910 2308 3084 1197 4942 D 1051 2983 1640 1291 2890 Whetstone 3720 5049 7574 5650 11450 "S" indicates single-precision, "D" indicates double-precision. Figures are in "kflops," thousand floating point operations per second. C:\EXOTIC\MATH If you are writing your own programs for in-house scientific use there are even more esoteric options. Computers based on the Inmos/SGS T800 transputer and the Intel i860 RISC CPU chip are available in the form of expansion boards that fit in your AT or 386 expansion slots. The T800 operates at 10 MIPS and 1.5 megaflops. The i860 is capable of from 8 to 20 megaflops. Microway offers these "supercomputers-on-a-board" and the software (C, Fortran, and utilities) to run them. A German company, DSM, phone 408-946- 0655, also offers an i860 computer on an expansion board. C:\RECAP If you need more number crunching speed: 1. If you don't have a numeric co-processor on your system, see if your software applications support one. If they do support one, consider adding a co-processor if you meet the compute-bound criteria previously described. 2. If you already have a co-processor and applications that use it, consider switching to a Cyrix, IIT, or Weitek (if your applications can also support the Weitek) before upgrading to a faster CPU. 3. If you are shopping for a new number crunching computer, look for: a high Mhz rating of the CPU, lots of fast motherboard memory, lots of memory cache (not just disk cache), and ESDI or SCSI disks with a 1:1 controller. Then shop for the fastest co- processor that both the motherboard and your applications support. FidoNews 7-52 Page 8 24 Dec 1990 4. Beware of the claims of the 486-25. It's not as fast as a 386-33 with a Weitek 3167. The extra cost is not in proportion to the extra power. Aside from the FPU, it's no faster than a vanilla 386-33. But if you have the bucks to spend, the 486 combined with a Weitek 4167 is the current ultimate. 5. To shop for a co-processor check the chip distributors who advertise in the back of magazines such as Byte and PC Magazine. When you are ready to buy, be sure to tell the vendor what kind of system you have in order to match the part # and Mhz rating of the chip. 6. A company specializing in co-processors, add-in boards, and number crunching, both software and hardware, is Microway, P.O. Box 79, Kingston, Mass. 02364. (508)746-7341 7. Other chip distributors are: JDR Microdevices 1-800-538-5000 I.C. Express 1-800-877-8188 R&R Electronics 1-800-736-3644 Microprocessors Unl. 1-918-267-4961 Ann & Anthony 1-408-988-5083 (These are just for reference. I'm not endorsing any.) [end] ----------------------------------------------------------------- FidoNews 7-52 Page 9 24 Dec 1990 AVICULTURISTS WANTED!! I have noticed reading the echo's available to me in Region 12, that there is not one dedicated to Aviculturists. For those that are unfamiliar to this term, it relates to those interested in the breeding of domestivated birds (ie Canaries, Parrots etc etc). I've seen echo's devoted to those interested in Tropical fish (the AQUARIUM echo) and thought of establishing one for those involved in the breeding or just the keeping of birds. If there is one already established, I would most certainly like to hear about it. If not, I would appreciate hearing from anyone that is interested in such an echo. If demand prevails, I am considering setting up an echo area. Any Sysops who are interested, either for themselves or for their users, can contact me via Net Mail, 24 hrs, at 1:229/110 Durham Systems (Online). Paul Chantler Durham Systems (Online) 1:229/110 Hopeful future host of the AVIARY echo ----------------------------------------------------------------- FidoNews 7-52 Page 10 24 Dec 1990 Chris Mayer, Ghostbusters Central (415) 933-4937 FidoNET 1:161/303 - GateNET 84:4101/303 The GhostNET Echo Yes, another echo is out, and its called 'GhostNET'. This new echo is a serious attempt at disscussing topics relating to paranormal occurences, Extrasensory Perception, Psychokinesis, Psychics, and more. Now that I've got you confused, here are the basic requirements. It requires real names and ANSI codes are to be stripped. To recieve this echo, send netmail to one of the following nodes: (Please choose the node closest to you) Chris Mayer @ 1:161/303 ------------------ 415-933-4937 Chris Mayer @ 84:4101/303 (From GateNET) - 415-933-4937 Cheryl Mathison @ 1:161/69 ------------------- 415-481-2806 Todd Looney @ 1:143/27 ------------------- 408-298-2740 Dan Hyman @ 1:205/45 ------------------- 209-661-5355 Dolores Jensen @ 1:142/999 ------------------ 203-742-7205 ----------------------------------------------------------------- FidoNews 7-52 Page 11 24 Dec 1990 Robert Keown FidoNet 1:273/718 RbbsNet 8:952/8 PUBNET: The International Publisher's Exchange For those interested in electronic publishing, desktop publishing, and computer graphics, a new organization has been formed with telecommunications technology at its core. The International Publisher's Exchange (PUBNET) is a not-for-profit organization which serves as an information clearing house for individuals involved in graphic communications. Currently, PUBNET is carried by 200 BBS's throughout the U.S. and Europe. Primarily implemented with Fido-based technology, PUBNET is expanding its operations to include gateways to other networks (i.e. InterNet & RIME). PUBNET currently consists of 14 echoes focused on many aspects of computer graphics and publishing including specific software topics (i.e. PageMaker, Ventura, Quark, etc.), production techniques, hardware technologies, and other such forums. Additionally, PUBNET maintains a file distribution system as well as a specific echo, called PUBNEWS, which carries late-breaking news and press releases regarding the fast-moving graphic communications market. Many times new product announcements appear in PUBNEWS long before they appear in print. Users of PUBNET include those with a casual interest or need, to professional individuals and organizations involved in graphic production, computer software and hardware, and corporate communications. PUBNET provides a good way for all levels of users to learn and share ideas and solutions in computer publishing. If you would to learn more about PUBNET, or would like to establish echo feeds, you may contact Carl Dickson, 1:109/519, Robert Keown, 1:273/718, or Gene Rodrigues, 1:143/25, for additional information and the location of the nearest PUBNET hub. ----------------------------------------------------------------- FidoNews 7-52 Page 12 24 Dec 1990 Michael Rapp FidoNet 1:106/12 /------------------------------\ | The Respectable SysOps Guide | | by: Michael Rapp | LAST REVISED: 12/17/90 | aka: Captain Vizor | | SysOp: USS Vulcanix | | 713/774-0088 | \------------------------------/ This is a guide that ALL SysOps should follow if they are to be viewed by the modem world as "Respectable", and their BBSes known a "Quality" BBS. ----------------------------------------------------------- 1. Respectable SysOps don't put up a board without a hard drive. They already have a 20MB or greater one made by a major company like Seagate. 2. Respectable SysOps have 2400 baud or greater. The one exception is that chat boards may have 300 because of the expense of 5 or more phone lines and modems. 3. Respectable SysOps do not run their BBS when the fan has broken down. 4. Respectable SysOps turn off their monitor as often as possible. 5. Respectable SysOps don't care if a user is going to put up a board next week. 6. Respectable SysOps try to help the user get his/her board started by offering hints and tips. 7. Respectable SysOps who operate boards that require money for access do not give free validation to someone just because they run "THE ULTIMATE BBS" and have a 250 MB hard drive. 8. Respectable SysOps laugh when users say 'I have your voice number' or 'I am a Phed'. 9. Respectable SysOps are very annoyed when users leave numbers like 'CAN-NOT-TELL' or 'PRI-VAT-E!!!'. 10. Respectable SysOps do not make up excuses to leave a person with whom they are chatting like 'My dad is calling me'. FidoNews 7-52 Page 13 24 Dec 1990 11. Respectable SysOps NEVER have pirated software on their boards. 12. Respectable SysOps ALWAYS check uploaded software for viruses before putting it online. 13. Respectable SysOps have a dedicated computer with a dedicated phone line and modem for their board. They never take it down except to run BBS maintinence or to call a few BBSes that they like. They never tie up the board for more than an hour calling other BBS systems. 14. Respectable SysOps don't say 'L8r', 'k-K00l','B@SS' or any other related terms. 15. Respectable SysOps when on chat boards do not use macros every five minutes saying "Call the SPECTRUM ELITE! ???/???-????". 16. Respectable SysOps NEVER leave mail to users asking them to post. 17. Respectable SysOps get EXTREMELY angry if their boards are crashed intentionally by other users. If unintentionally, they go get a Pepsi and try to fix it. 18. Respectable SysOps ALWAYS make a backup of their user list and other important files every week. 19. Respectable SysOps don't care about "improper sign-offs". 20. Respectable SysOps don't think that they are God and are better than everyone else. 21. Respectable SysOps help the new users, not cut them down. 22. Respectable SysOps NEVER have a message base such as "ABUSE THAT USER". 23. Respectable SysOps don't care what another user posts about him/her on a "kid" or "looser" or "sleazy" board. 24. Respectable SysOps put up the BBS as a service. 25. Real SysOps don't get out their sector editor and put their name on every piece of software they have in hopes of gaining national recognition. 26. Respectable SysOps NEVER let users threaten them in any way. FidoNews 7-52 Page 14 24 Dec 1990 27. Respectable SysOps NEVER EVER answer their board line by voice. 28. Respectable SysOps NEVER chat with users when they themselves are in a bad mood. 29. Respectable SysOps know that users tend to like SysOps who take the time to chat to work out a problem the user is having online. 30. Respectable SysOps always wait at least 15 seconds before breaking into chat. 32. Respectable SysOps never type their fastest. Users respect SysOps who are accurate. 33. Respectable SysOps NEVER say "My board is the best" or "if you're going to pay, make it my board", this will more discourage users than attract them. 34. Respectable SysOps NEVER copy an obvious, original idea from another board. Users admire SysOp that come up with their own ideas. 35. Respectable SysOps try to modify the original BBS software as much as possible. Users tend to get bored if they call boards that just leave the software exactly in the original "generic" condition. Well....there you go! My first submission to FidoNews! If you like it, or if you would like to add anything to it, just send me netmail! Thanks. ----------------------------------------------------------------- FidoNews 7-52 Page 15 24 Dec 1990 NewStyle Packets A Proposal for the Next Generation of of FidoNet Mail Packers Fourth Draft 19 December 1990 jim nutt 1:114/30@fidonet Introduction FidoNet has been using the Type II style packet for some five years or more now with good results. However, at this point, the Type II format has been extended an amazing number of ways using the "Kludge" hidden line facility provided by a leading ^A (ASCII SOH) on a line of text. It is my belief that the time has come to move to a newer technology for handling packets, one that is inherently extensible and easily handled by a number of systems. Such a system should be able to handle such varied things as integrated text/graphics and other special attributes of messages. Basic Format Essentially, this format would break a message into a number of "chunks". Each chunk would be a maximum of 4,294,967,306(!) bytes long including its header and may contain any type of data. A chunk header would be 21 bytes long and would consist of a 4 byte chunk type tag followed by an 8 byte length field. The length field does *not* include the 12 bytes of the chunk header. Additionally, the length field is the the length of the chunk *AS TRANSMITTED*, i.e., in ASCII. Chunks would be unterminated. In C, a chunk structure would look like this: struct chunk { char type[4]; char len[8]; /* 32 bit length of data field, 8 hex digits */ unsigned char data[len]; /* not really, this isn't legal c, but it gets the idea across */ }; Certain chunk types require that a FidoNet address be represented in a binary hex format. This address would be comprised of the domain, zone, net, node, and point expressed as the following C structure: struct address { FidoNews 7-52 Page 16 24 Dec 1990 char point[4]; char node[4]; char net[4]; char zone[4]; char domain[]; }; The domain name is nul terminated and variable length. If absent it defaults to the current domain. All other fields are 4 hex digits with NO terminating nul character. It was chosen to use an ASCII representation of numbers (in hex) to avoid byte ordering problems and to enhance portability across 7 bit transport layers. For example, the author's address would be: 0000001E00720001FIDONET See Appendix 6 for suggestions on generating this form of address. A line, as defined by this document, is a sequence of 7 bit ASCII characters terminated with a linefeed (0x0a). Chunk Types Chunk type names are exactly four characters long, padded with spaces if necessary. Chunk types not recognized by a program would be passed along and ignored. Chunk types that are marked with an asterisk (*) must be recognized by a conforming installation. Chunk types marked with a C are considered control chunks, while those marked with D are data chunks. Unmarked chunks are delimiters or informational. I would propose the following base chunk types: * BEGB A chunk indicating the beginning of a bundle. This chunk may contain optional information identifying the bundle. CRTR Indicates the software and revision level used to create this bundle. Applies only to entire bundles. * PSWD Password for the entire bundle, or if within "BEGM"/"ENDM" a single message. If the password in this chunk does not match a predefined password on the receiving system one of two actions occurs. If the receiving system is the final destination of the bundle or message, the bundle or message is discarded, optionally with a message being sent back to the sender saying so. FidoNews 7-52 Page 17 24 Dec 1990 If the bundle or message is only passing through, it will not be made visible to the sysop of the routing system, regardless of any options that may be set to the contrary. Obviously, this is lightweight security, but it is better than nothing! BEGA C This chunk indicates the conference the following messages (up to the corresponding ENDA marker) belong to. This could be multiple areas, one area per line. BEGA/ENDA chunks may be nested. * BEGM A chunk indicating the beginning of a message, this chunk may contain optional information identifying the message. * ROUT C Binary address of next destination for this message or bundle. In other words, if a message from 123/456 is going to 456/789 but is routed through an intermediary system (say 321/654) this address would be that of the intermediary system. This chunk may be applied to either a single message or an entire bundle. * TO C Name and address of receiver in ASCII. The address in this field may be anything, so long as the system at the "ROUT" address can make sense of it. For split addresses (name and address separated, such as a FidoNet address) the address should be on two ASCII lines separated a newline (ASCII 0x0a) with the address on the first line and name on the second line. Integrated addresses (name and address all in one, such as Internet) should be on one line, with an optional second line giving the recipient's actual name. * FROM C Name and address of sender in ASCII. This may be anything so long as it is possible for the receiver to reply via the address in this field. The format of this should be the same as for the TO chunk * TIME D Timestamp of the message. Represented as a 32 bit integer counting the number of seconds from January 1, 1970 0:00 UTC. 8 digit hexadecimal field. * SUBJ D Subject of the message FidoNews 7-52 Page 18 24 Dec 1990 * ATTR C Attributes of the message. See Appendix 2 for a complete list of message attributes. * NUMB Serial number of this message on originating system. This chunk is fixed as an 8 byte hex word. Length is 8. * RPLY Identifier of the message this is a reply to. This is a two line field, with the address of the system the message being replied to on the first line and the serial number (from the NUMB) chunk of the replied to message on the second line. * ATCH C Name of a file attached to this message * FREQ C Name of a file requested from receiving system. This would incorporate the same type of update request logic as is currently used by WaZoo mailers. A separate "FREQ" chunk is required for each file requested. DOMN C Echomail only, list of domains, one per line, of the domains this message has passed through ZONE C Echomail only, list, as four byte hex words, of zones that have seen this message. This chunk is cleared each time the message enters a different domain and the name of the domain the message is exiting is added to the "DOMN" chunk. NET C Echomail only, list, as four byte hex words, of all nets that have seen this message. This chunk is cleared upon export to another zone and the exporting node's zone number is added to the "ZONE" chunk. NODE C Echomail only, list, as four byte hex words, of all nodes in the current net that have seen this message. This chunk is cleared each time the message enters a new net and the number of the net the message is exiting is added to the "NET " chunk. PONT C Echomail only, list, as four byte hex words, of all point systems that have seen this message. This chunk is cleared upon export to another node and the node number of the exporting system is added to the "NODE" chunk. * PATH List of the systems this message has passed FidoNews 7-52 Page 19 24 Dec 1990 through to reach this system, in order. This includes all systems in all zones and domains. All addresses would be in hex as defined in the section "Basic Formats" with one address per line. * TEXT D The text of a message. This is restricted to seven bit ASCII characters. Prefixing a character with an ASCII ESC (0x1b) indicates that the following character has its high-order bit set. ASCII ESC is recognized by the sequence ESC ESC. Text lines are terminated with an ASCII LF (0x0a). A message may contain as many TEXT chunks as are necessary for the text of the message. Adjacent TEXT chunks are logically concatenated. QUOT C Indicates that the following text chunk is quoted. This field contains: name of quoted person address of quoted person timestamp of quoted message (8 digit hex) optional information with each subfield on a separate line. GRPH D A vector drawing. This is an HPGL image. Multiple GRPH chunks may appear in a message. GRPH chunks, like TEXT chunks, are logically concatenated when adjacent. SEEN C This is a list of the systems that have seen this message. Address are in ASCII, one per line. Omitted fields in an address are replaced by the value of the corresponding field of the preceding address. * ENDM A chunk indicating the end of a message. This chunk may optionally contain information identifying the message it terminates. ENDA This chunk terminates a list of message for a particular conference(s). If the opening BEGA chunk lists multiple conferences, then multiple ENDA chunks may be used, each listing one or more of the conferences in the BEGA. See Appendix 4 for more details. * ENDB A chunk indicating the end of the bundle, anything after this can be safely ignored. FidoNews 7-52 Page 20 24 Dec 1990 This chunk may optionally contain information identifying the bundle it terminates. NULL An empty chunk. This may never be needed, but would conceivably be useful for separating adjacent GRPH and TEXT chunks that should not be concatentated. Other Considerations Chunk style packets could not be sent as *.PKT files as they are not backward compatible with type II packets. I propose that chunk style packets be called bundles and sent as *.BUN files, with compressed bundles sent as *.B?? where ?? is the compression method used (see Appendix 1 for extensions). Bundle file names should be unique for at least a one week cycle, a 32 bit serial number expressed in hexadecimal should prove adequate for most applications. Experimental chunk types are provided for by the provision that unrecognized chunk types be passed through and ignored. Systems that know how to use a particular chunk type (say, BITS) can, while systems that don't understand it may ignore it. Chunks should appear in a bundle in roughly the same order as they appear above, with control and informational chunks (PATH, ROUT, etc) appearing before data chunks (TEXT, BITS, GRPH). Control Chunk tag name assignments are controlled by Appendix 3 of this document. New chunk tags may be added and old ones revised by revision of this document. Message attribute assignments are controlled by Appendix 2 of this document. New attributes may be assigned by revision of this document. Bundle file extensions are controlled by Appendix 1 of this document. New extensions may be defined and old ones revised by revision of this document. Conclusion I doubt I have covered all possible or desirable chunk types in this document. I do believe however, that enough have been defined to get started with. Chunks offer a highly flexible, extensible system of bundling mail. New types of chunks may defined as needed to accomodate advances in technology and FidoNet. Additionally, this would further separate the application and transport layers of FidoNet, yielding less confusion as to their respective roles. FidoNews 7-52 Page 21 24 Dec 1990 It may be noticed that this structure is extremely similar to the IFF format as used on Amiga computers and introduced by Electronic Arts Software. While inspired by IFF, this system has been simplified somewhat and changed to be more easily transportable between computers using different byte orders and processors. All fields defined in this document are 7 bit ASCII and should be easily parsed by any system. Appendix 1 - Compression Extensions Compressed bundles would indicate the type of compression used by the following file extensions: Extension Creator --------- ------- .BUN Uncompressed .BPK PKZip .BLH LHarc .BAR ARC .BDW DWC .BPA PAK .BZO ZOO .BPX PKXarc Appendix 2 - Message Attributes The following attributes have been assigned: PRIV privileged message (sysop or recipient only) CRYP encrypted message CRSH High priority message DRCT Send directly to destination (no routing) HOLD Hold this message for pickup Attributes are listed one per line and may combined where it makes sense to do so. (crash and hold makes no sense...) Appendix 3 - Defined Chunk Tags The following chunk tags are defined in this document: BEGB TO RPLY NODE SEEN CRTR FROM ATCH PONT ENDM PSWD TIME FREQ PATH ENDA BEGA SUBJ DOMN TEXT ENDB BEGM ATTR ZONE QUOT NULL ROUT NUMB NET GRPH Appendix 4 - The BEGA/ENDA chunks FidoNews 7-52 Page 22 24 Dec 1990 Messages may belong to multiple conferences (this is quite common in UseNet). Therefore it is possible to nest both BEGA and ENDA chunks. Additionally, it is possible to list multiple conferences in either chunk. For example, ten messages come through a system. The first five are crossposted to both c_echo and c_plusplus, the next two are c_echo only and the next are posted to c_echo and net_dev. Finally, the last message is to net_dev only. The basic bundle structure (showing only the BEGA/ENDA [with BEGM....ENDM indicating a message] chunks for clarity) would be as follows ( is a linefeed [ASCII 10]): BEGA00000011C_ECHOC_PLUSPLUS BEGM....ENDM BEGM....ENDM BEGM....ENDM BEGM....ENDM BEGM....ENDM ENDA0000000AC_PLUSPLUS BEGM....ENDM BEGM....ENDM BEGA00000007NET_DEV BEGM....ENDM BEGM....ENDM ENDA00000006C_ECHO BEGM....ENDM ENDA00000007NET_DEV In other words, BEGA/ENDA are cumulative. This should allow greater flexibility in conferencing and should assist in porting UseNet conferences over as it is common for a UseNet message to posted across several conferences. Appendix 5 - Echomail This proposal contains two different possibilities for controlling echomail flow. The first assumes a tree structured topology and is represented by the DOMN, ZONE, NET, NODE and PONT chunks. The second is similar to current echomail distribution and uses the SEEN chunk to list the systems a message has been seen by. This document does not attempt to judge either system on its merits (or otherwise), they are simply presented. Appendix 6 - Addresses and C In C, chunk addresses can be generated using the sprintf() function with parameters of the form: sprintf(addrstr,"%04x%04x%04x%04x%s",point,node,net,zone,domain); It may be read back with: FidoNews 7-52 Page 23 24 Dec 1990 sscanf(addrstr,"%4x%4x%4x%4x%s",&point,&node,&net,&zone,&domain); Appendix 7 - Contributors The following people have contributed to this specification in some way. If I've missed your name, please forgive me... I get a lot of mail on this and have misplaced some messages. Alastair Rankine of 3:680/820@fidonet Alexander Holy of 2:310/11.32@fidonet Andreas Kaiser of 2:509/5@fidonet.org Andy Lester of 1:115/934@fidonet Baba Buehler of 1:233/4.0@fidonet Brendan Morley of 3:640/297.3@fidonet Charles Falconer of 1:141/209.1@fidonet Clinton Adams of 1:273/101@fidonet Daniel Kalchev of 2:359/1@fidonet David Nugent of 3:632/348.65535@fidonet Erik Seielstad of 1:260/231@fidonet Fabian Gordon of 1:107/323@fidonet Garth Kidd of 3:680/828@fidonet Gerard van.der.Land of 2:283/1.5@fidonet Henk Wevers of 2:500/1@fidonet Ian McLaughlin of 1:353/250@fidonet.org Jan Terpstra of 2:512/10.0@fidonet Jan Vroonhof of 2:281/1.12@fidonet Jeffrey Nonken of 1:273/715.2@fidonet Joe Jared of 1:103/201.1@fidonet Johan Zwiekhorst of 2:292/100@fidonet Juergen Hermann of 2:507/642.7610@fidonet Mark Howard of 1:260/340.1@fidonet Maximilian Hantsch of 2:310/6@fidonet Pieter Muller of 5:492/1.4@fidonet Renald Loignon of 1:167/176.1@fidonet Rob Tillotson of 1:201/40.302@fidonet Robert Heller of 1:321/153.0@fidonet Roger Mordin of 2:201/231.10@fidonet Ronald Bruintjes of 2:281/600@fidonet Russell McOrmond of 1:163/109.1@fidonet Sico Bruins of 2:512/10.4@fidonet Torben Paving of 2:231/16.9@fidonet FidoNews 7-52 Page 24 24 Dec 1990 ----------------------------------------------------------------- FidoNews 7-52 Page 25 24 Dec 1990 ================================================================= COLUMNS ================================================================= "Captain's Log, Stardate 9012.25..." by Phil Buonomo, 1:107/583@FidoNet 520/583@AlterNet 9:807/1@PNet Since this is the last scheduled FidoNews of the year, and coin- cidentally the last one before BOTH Christmas and New Year's, I made a special effort of submitting something this week. I've a few articles that have been on the back burner for a few weeks, but just haven't had the time to get them in. Hopefully I'll be able to write them down soon. In the meantime I wish all who read this a very happy holiday season, regardless of religious belief, color of skin, sexual preference, or even network affiliation: A Christmas Poem (Author unknown) Twas the night before Christmas, and all through the shop, The computers were whirring; they never do stop. The power was on and the temperature right, In hopes that the input would feed back that night. The system was ready, the program was coded, And memory drums had been carefully loaded; While adding a Christmasy glow to the scene, The lights on the console flashed red, white, and green. When out in the hall there arose such a clatter, The programmer ran out to see what was the matter. Away to the hallway he flew like a flash, Forgetting his key in his curious dash. He stood in the hallway and looked all about, When the door slammed behind him,and he was locked out. Then, in the computer room, what should appear, But a miniature sleigh and eight tiny reindeer; And a little old man, who with scarcely a pause, Chuckled: "My name is Santa... the last name is Claus." The computer was startled, confused by the name, Then it buzzed as it heard the old fellow exclaim: "This is Dasher and Dancer and Prancer and Vixen, And Comet and Cupid and Donner and Blitzen." With all these odd names, it was puzzled anew; It hummed and it clanked, and a main ciruit blew. It searched in its memory core, trying to "think"; Then the multi-line printer went out on the blink. Unable to do its electronic job, It said in a voice that was almost a sob: "Your eyes--how they twinkle--your dimples so merry, Your cheeks so like roses, your nose like a cherry, Your smile--all these things, I've been programmed to know, And at data-recall, I am more than so-so; FidoNews 7-52 Page 26 24 Dec 1990 But your name and your address (computers can't lie), Are things that I just cannot identify. You've a jolly old face and a little round belly, That shakes when you laugh like a bowlful of jelly; My scanners can see you, but still I insist, Since you're not in my program, you cannot exist!' Old Santa just chuckled a merry "ho,ho," And sat down to type out a quick word or so. The keyboard clack-clattered, its sound sharp and clean, As Santa fed this "data" into the machine: "Kids everywhere know me; I come every year; The presents I bring add to everyone's cheer; But you won't get a thing-- that's plain to see; Too bad your programmers forgot about me." Then he faced the machine and said with a shrug, "Happy Christmas to all," as he pulled out its plug. ----------------------------------------------------------------- FidoNews 7-52 Page 27 24 Dec 1990 ================================================================= LATEST VERSIONS ================================================================= Latest Software Versions MS-DOS Systems -------------- Bulletin Board Software Name Version Name Version Name Version DMG 2.93 Phoenix 1.3 TAG 2.5g Fido 12s+ QuickBBS 2.66 TBBS 2.1 GSBBS 3.01 RBBS 17.3A TComm/TCommNet 3.4 Lynx 1.30 RBBSmail 17.3B Telegard 2.5 Kitten 2.16 RemoteAccess 0.04a TPBoard 6.1 Maximus 1.02 SLBBS 1.77A Wildcat! 2.50 Opus 1.13+ Socrates 1.10 WWIV 4.12 PCBoard 14.5 XBBS 1.15 Network Node List Other Mailers Version Utilities Version Utilities Version BinkleyTerm 2.40 EditNL 4.00 ARC 7.0 D'Bridge 1.30 MakeNL 2.31 ARCAsim 2.30 Dutchie 2.90C ParseList 1.30 ARCmail 2.07 FrontDoor 1.99c Prune 1.40 ConfMail 4.00 PRENM 1.47 SysNL 3.14 Crossnet v1.5 SEAdog 4.51b XlatList 2.90 DOMAIN 1.42 TIMS 1.0(Mod8) XlaxDiff 2.35 EMM 2.02 XlaxNode 2.35 Gmail 2.05 GROUP 2.16 GUS 1.30 HeadEdit 1.15 InterPCB 1.31 LHARC 1.13 MSG 4.1 MSGED 2.00 MSGTOSS 1.3 PK[UN]ZIP 1.10 QM 1.0 QSORT 4.03 Sirius 1.0x SLMAIL 1.36 StarLink 1.01 TagMail 2.41 TCOMMail 2.2 Telemail 1.27 TMail 1.15 TPBNetEd 3.2 FidoNews 7-52 Page 28 24 Dec 1990 TosScan 1.00 UFGATE 1.03 XRS 4.00* XST 2.2 ZmailQ 1.12 OS/2 Systems ------------ Bulletin Board Software Network Mailers Other Utilities Name Version Name Version Name Version Maximus-CBCS 1.02 BinkleyTerm 2.40 Parselst 1.32 ConfMail 4.00 EchoStat 6.0 oMMM 1.52 Omail 3.1 MsgEd 2.00 MsgLink 1.0C MsgNum 4.14 LH2 0.50 PK[UN]ZIP 1.02 ARC2 6.00 PolyXARC 2.00 Qsort 2.1 Raid 1.0 Remapper 1.2 Tick 2.0 VPurge 2.07 Xenix/Unix ---------- BBS Software Mailers Other Utilities Name Version Name Version Name Version MaximusCBCS 1.02.Unix.B0 BinkleyTerm 2.30b Unzip 3.10 ARC 5.21 ParseLst 1.30b ConfMail 3.31b Ommm 1.40b Msged 1.99b Zoo 2.01 C-Lharc 1.00 Omail 1.00b FidoNews 7-52 Page 29 24 Dec 1990 Apple CP/M ---------- Bulletin Board Software Network Mailers Other Utilities Name Version Name Version Name Version Daisy v2j Daisy Mailer 0.38 Nodecomp 0.37 MsgUtil 2.5 PackUser v4 Filer v2-D UNARC.COM 1.20 Macintosh --------- Bulletin Board Software Network Mailers Other Utilities Name Version Name Version Name Version Red Ryder Host 2.1 Tabby 2.2 MacArc 0.04 Mansion 7.15 Copernicus 1.0 ArcMac 1.3 WWIV (Mac) 3.0 LHArc 0.33 Hermes 1.01 StuffIt Classic 1.6 FBBS 0.91 Compactor 1.21 TImport 1.92 TExport 1.92 Timestamp 1.6 Tset 1.3 Import 3.2 Export 3.21 Sundial 3.2 PreStamp 3.2 OriginatorII 2.0 AreaFix 1.6 Mantissa 3.21 Zenith 1.5 Eventmeister 1.0 TSort 1.0 Mehitable 2.0 UNZIP 1.02c Amiga ----- Bulletin Board Software Network Mailers Other Utilities Name Version Name Version Name Version Paragon 2.07+ BinkleyTerm 1.00 AmigArc 0.23 TransAmiga 1.04 TrapDoor 1.50 AReceipt 1.5 FidoNews 7-52 Page 30 24 Dec 1990 WelMat 0.42 booz 1.01 ConfMail 1.10 ChameleonEdit 0.10 ElectricHerald1.66 Lharc 1.21 MessageFilter 1.52 oMMM 1.49b ParseLst 1.30 PkAX 1.00 PK[UN]ZIP 1.01 PolyxAmy 2.02 RMB 1.30 Skyparse 2.30 TrapList 1.12 UNzip 0.86 Yuck! 1.61 Zoo 2.01 Atari ST -------- Bulletin Board Software Network Mailer Other Utilities Name Version Name Version Name Version FIDOdoor/ST 2.00* BinkleyTerm 2.40j* ConfMail 4.02 Pandora BBS 2.41c The BOX 1.30 ParseList 1.30 QuickBBS/ST 1.02 ARC 6.02 GS Point 0.61 FiFo 2.1c LHARC 0.60 LED ST 0.10 BYE 0.25 PKUNZIP 1.10 MSGED 2.01 SRENUM 6.2 Trenum 0.10 OMMM 1.40 Archimedes ---------- BBS Software Mailers Utilities Name Version Name Version Name Version ARCbbs 1.44 BinkleyTerm 2.03 Unzip 2.1TH ARC 1.03 !Spark 2.00d FidoNews 7-52 Page 31 24 Dec 1990 ParseLst 1.30 BatchPacker 1.00 + Netmail capable (does not require additional mailer software) * Recently changed Utility authors: Please help keep this list up to date by reporting new versions to 1:1/1. It is not our intent to list all utilities here, only those which verge on necessity. ----------------------------------------------------------------- FidoNews 7-52 Page 32 24 Dec 1990 ================================================================= NOTICES ================================================================= =========== >>> PERNnet: Brigadoon Weyr <<< ========== *********** >>> Gift Certificate <<< ********** Welcome to the planet PERN where one can play Pernese style of life. With this Certificate you may get ONE link into PERNnet regardless of the originating system's origin and network. Brigadoon Weyr is devoted to the weirder things in life -- fringe groups, nonstandard religions, Elvis searchers, cult movies, bad audio and video, and certified maniacs and politicians. It is the goal of The Brigadoon Weyr to allow the weird, the odd, the huddled masses yearning to be free an electronic playpen in which to espouse their varied belief systems. It is a network of losers, winners, dreamers, screamers. In time the Pernese conference where one will play out the roll of a Pernese name and also be able to have (a) Fire Lazard(s) or be a dragon rider to defend the planet Pern againsts the Red Planet that visits every 400 years. In closing, I point out that we are all of us doing this as a hobby, and will tolerate no politics or powerplays. We make no distinctions as to age, background, religion, sex, gender, et cetera. We are an Equal Opportunity Network, limited solely by the ability to move echomail around in a reasonably good manner. Problems with offending nodes will be dealt with via the path structure of the network until it reaches The Brigadoon Weyr Central (1:325/101@FidoNet) where we will deal with it as need be. For your free information package you may file requet(FREQ) the magic name of BRIGADOON from the following locations: Also you may FREQ the magic name of PERNLIST and it will get you a list of systems currently part of PERNnet. 1:325/101 @FidoNet 9:1010/11 @ParaNet 69:2700/1 @AdultLinks(or KinkNet) also available at: 1:363/71 @FidoNet 3:712/634 @FidoNet This should get you BRIGADON.ZIP. if you need a standard ARC version FREQ `BRIGADON.ARC' instead. FidoNews 7-52 Page 33 24 Dec 1990 ----------------------------------------------------------------- The Interrupt Stack 1 Jan 1991 Implementation of 7% Goods and Services Tax in Canada. Contact Joe Lindstrom at 1:134/55 for a more colorful description. 16 Feb 1991 Fifth anniversary of the introduction of Echomail, by Jeff Rush. 31 Mar 1991 Jim Grubs (W8GRT) was issued his first ham radio license forty years ago today. His first station was made from an ARC-5 "Command Set" removed from a B-17 bomber. 12 May 1991 Fourth anniversary of FidoNet operations in Latin America and second anniversary of the creation of Zone-4. 8 Sep 1991 25th anniversary of first airing of Star Trek on NBC! 7 Oct 1991 Area code 415 fragments. Alameda and Contra Costa Counties will begin using area code 510. This includes Oakland, Concord, Berkeley and Hayward. San Francisco, San Mateo, Marin, parts of Santa Clara County, and the San Francisco Bay Islands will retain area code 415. 1 Feb 1992 Area code 213 fragments. Western, coastal, southern and eastern portions of Los Angeles County will begin using area code 310. This includes Los Angeles International Airport, West Los Angeles, San Pedro and Whittier. Downtown Los Angeles and surrounding communities (such as Hollywood and Montebello) will retain area code 213. 1 Dec 1993 Tenth anniversary of Fido Version 1 release. 5 Jun 1997 David Dodell's 40th Birthday If you have something which you would like to see on this calendar, please send a message to FidoNet node 1:1/1. ----------------------------------------------------------------- FidoNews 7-52 Page 34 24 Dec 1990 Greylock Software is seeking testers for for a multiple host point system. If you are interested, particularly if you use other mail systems such as MCI and CIS in addition to FidoNet, send netmail to 1:321/202@fidonet. ----------------------------------------------------------------- Michael Rapp FidoNet 1:106/12 Notice To All Star Trek SysOps -- Looking for some new Star Trek Files? Why don't you FREQ the magic file name "FILES" from my board and pick out the ones you want! You can FREQ a maximum of five files per session and 250KB per session. Please only FREQ once per day. Thank you. Also, I run a BBS list of Star Trek Related BBSes and I distribute it regularly around Houston (I also make it available for FREQ, "STBBS.LST"). If you would like to be added to this list, please send me netmail with the following information: 1. BBS Name 2. FidoNet Address and any other addresses you have (please specify net name) 3. SysOp Name (Handle if desired) 4. Board Phone (with area code!) 5. Maximun Baud Rate 6. BBS Software and Version Number 7. Hours if part-time BBS Thanks! -----------------------------------------------------------------