Ber# TR Lengte Aan Van @ BBS Datum/tijd Titel 658 BN 5962 ALL ON7KL RCNET 0625/0743 Netrom v Thenet another view R:880623/0651 @:PI8AYB #2083 [Forwarding - PI8AYB - Putten - R.34] R:880623/0619l @:PA3APN #1441 [PA3APN MAARSSEN. Wide Area Network.#12] R:880623/0534L @:PE1AUE #5161 [via PE1AUE - EURO-PA #3 * CM54f-JO22hi> R:880620/0316z 1482@PA0QRB [Vlaardingen JO21EW] R:880619/2004z 5128@PA3EAE * Naaldwijk - Westland * R:880619/0114 @:PI8HWB [430.675-144.650-50.400-3.605] #:8939 Z:BREDA R:880618/1051 @:ON1AWB Antwerp, Belgium #:1826 Z:B-2008 R:880618/0703 @:ON7RC JO20EU, Brussels, Belgium #:820 Z:B-1040 16 June 1988 Net/Rom versus The Net An independent report from Ronald R. McCallister, N7FYA When THE NET first appeared, I was concerned that my investment in Net/Rom from Software 2000 was in danger. I also was curious about The NET because research into new products is my livelyhood. I read all the comments from those at Software 2000 about piracy. I do NOT believe in stealing any program. I think that software programmers should be able to make a living because I am one. I realize that this report will make a lot of individuals mad at me but I am in the consulting business and my opinions are as objective and accurate as I am able to make them. Please feel free to make comments back to me at N7HFZ BBS here in Washington or mail to AI Research, Inc. P.O. Box 97044, Tacoma, Wa 98497. This report is broken into two major sections: 1. comparison of code and my opinion of the comparison. 2. personal observations concerning NET/ROM and THE NET. Section 1. I called Ron, WA8DED at Software 2000 to get permission to disassemble NET/ROM (N7FYA-8) for the purpose of comparing the disassembled code to the disassembled code of THE NET. I was given verbal permission to do so provided I destroyed all papers upon completion. I have done so. I disassembled the NET/ROM and THE NET using SLR Z80DIS. I found that the two products are about 85% identical. Since both products were compiled by two versions of one compiler and used the same libraries, I expected 60 to 65% of the code would be the same. This is normal in programming. When I talked to Ron @ Software 2000, he said that there were assembly code sections that had been hand massaged to improve performance but he failed to tell me which section. In assembly code on the Z80 there are only a few ways to do certain items efficently. This means that any two GOOD programmers working on different programs in Z80 code are likely to code in a similiar if not exact fashion. The names in the procedures will be different in the source code but will look the same in the object code. In 'C' there are many ways to code anything BUT to be efficent in the Z80 environment, you must optimize to the hilt. That means if you are trying to do a connect sequence in a TAPR type TNC2 and want to stay compatible with the rest of the amateur community, You mus follow a specific set of rules. These rules will make 80% of connect sequence code identical. As Ron Raikes said, the code in the two roms are very much alike. As to being identical...... NO WAY! THE NET has some distinct differences that make it the better of the two node controllers. 1. It can operate in a full duplex mode whereas NET/ROM cannot. 2. THE NET is considerably faster in its reponse to changing network conditions. This alone tells me that the code is better optimized. 3. There are numerous features in THE NET that the NET/ROM is incapable of doing because of the Call Encrytion code. 4. It also will not crash. I have tried to crash it and NET/ROM for 15 days. THE NET has better error handlers than the NET/ROM. I cannot give any more specifics than this because I would be giving away the code from NET/ROM. Section 2. Personal Observations. As a software programmer, I can see the need to make money and to provide a good income for my family. As a end user however, I cannot but wonder why I need to pay a company $100.00 for a pair of eproms with a program. The eproms cost in single quantity cost $7.00 each (150ns). Then you add programming time. And finally you add the cost of the software and manual. Now it sounds fair. Let's go buy the orginal NET/ROMs for a hilltop. The two nodes just cost us $100.00 with only ONE manual. Version 1.0. Ahhhhh. Version 1.2 just became available and it fixes a few of the bugs in version 1.0. What? You mean I have to send my two original eproms back to be reprogrammed and it costs me $35.00 for each. That's $70.00 It is now 5 months later and a new version is out. Version 1.3 Here we are again sending $35.00 per eprom to have the bugs removed. Now we have decided to change the SSID on the node for compatability. That is another $35.00 per eprom. I have now spent $310.00 for just the eproms and STILL there are bugs in the programming. I also only have one manual. I charge $35.00 an hour in my job. How much is Software 2000 paying their people to reprogram one eprom? If you look at the big companies like Microsoft, Micropro, and Borland you find that they send you updates of their software with brand new rewritten LARGE and multiple manuals for $25.00 to $35.00 . What does Software 2000 offer that makes that big a difference? Also, these other big companies offer a FREE upgrade if they fix bugs within a short period of time, usually 3 months. Does Software 2000? This NOT the end of my opinions but I will stop. My personal opinions are separate from my findings of the investigation into the code of NET/ROM and THE NET. I would like to see the original code of NET/ROM and compare it to THE NET source code I recently received. Without the NET/ROM orginial code to compare, I must say that I prefer THE NET in performance. Last but most important! I say thanks to Software 2000 for their contributions to Amateur Radio packet but I would to caution them from alienating those that give them their sales. Sincerely, Ronald R. McCallister, N7FYA