Path: ...!Xl.tags.giganews.com!local-3.nntp.ord.giganews.com!nntp.earthlink.com!news.earthlink.com.POSTED!not-for-mail NNTP-Posting-Date: Wed, 12 Feb 2025 21:24:37 +0000 Subject: Re: GIMP 3.0.0-RC1 Newsgroups: comp.os.linux.misc,comp.os.linux.advocacy References: <655acbf6-05e5-69ff-8a44-9f7075aafa2e@example.net> <9RycneStTKjZETf6nZ2dnZfqn_udnZ2d@earthlink.com> <73idnXWOHfZXYTb6nZ2dnZfqnPqdnZ2d@earthlink.com> From: "WokieSux282@ud0s4.net" Organization: WokieSux Date: Wed, 12 Feb 2025 16:24:40 -0500 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.13.0 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8; format=flowed Content-Language: en-US Content-Transfer-Encoding: 7bit Message-ID: Lines: 74 X-Usenet-Provider: http://www.giganews.com NNTP-Posting-Host: 99.101.150.97 X-Trace: sv3-5GqRGqhXCPCFxDXD3bK1S7n3f26ODt0M9udJLabIkmvbN0v+WMGiGWUxtCZqzPJAKEqK+04vD/oCdgW!SsxewEeFo8dM5LiIDVM+G+zbFHS4FmCjgRcFdUwwtJIMzwkxqp+3u7rDsbVZrHUJSWVw5/iyxKxF!RxEcjJRE3Tj+I+YVX0p6 X-Abuse-and-DMCA-Info: Please be sure to forward a copy of ALL headers X-Abuse-and-DMCA-Info: Otherwise we will be unable to process your complaint properly X-Postfilter: 1.3.40 Bytes: 5568 On 2/12/25 1:50 PM, rbowman wrote: > On Wed, 12 Feb 2025 12:25:17 -0500, WokieSux282@ud0s4.net wrote: > >> I think the 86 series had 'more future possibilities' >> than the 8085. There were too many 8-bit systems out there already, >> so bumping up to 16 bit was smart for sales. Why make/compete-with >> "just another TRS-80" ? > > 16/32 bit processors were in the air so it would make no sense to stay > with 8 bits. > > https://spectrum.ieee.org/the-inside-story-of-texas-instruments-biggest- > blunder-the-tms9900-microprocessor > > That covers the ground from a slightly different perspective, a TI > engineer. It's interesting to speculate on IBM's view of future > possibilities. A large part of the company didn't think there was a > future. Intel thought the 432 was the future but that fell on its face. > Using the 8088 solved the peripherals problem but it also meant the > performance wasn't better than a Z80. Z80 designs were already doing bank > switching. The 8088 just had the additional registers to implement it. > > The TMS9900 wasn't a bad chip, if a little odd if you came from the Intel/ > Zilog world. I worked with it on one project. Because of TI's roots they > had a rad hard version > > https://retrocomputingforum.com/t/the-texas-instruments-tms-9900- > microprocessor/1370 > > That's a good description of the oddities. I did a little ASM for the 9900. It *was* a bit 'odd' however. Register-sets-in-RAM was unique, but it facilitated multi-user. Everybody remembers the PC, but the real intended use of the chip was in the 990 mini-computer, where multi-user/tasking was a must. Actual hardware support for that, weird. Branch and Load Workspace Pointer ....... Anyway, TI flubbed it and another early 16-bitter went down the toilet. There were a LOT of 'concept chips' back around the same time when 16+ bits was becoming real. Everyone was sure they had the Better Idea. Now, you might maybe find one in a surplus parts bin. (had a dream once about finding a TMS9900 - white with gold pins - in a surplus bin) > The first article points out the IBM was big-endian and suddenly thy were > transported into the little-endian world. Our legacy software uses ONC-RPC > which handle the byte order. Originally the system ran on RS6000 machines > where the reshuffling was a NOOP. As we started using Linux in house for > development, the x86 machines had to reverse the canonical big-endian > data. No problem. Then our clients moved to Windows while we still used > Linux leading to the absurdity of dual processing to move little-endian to > big-endian and back to little-endian. > > All that is hidden in the RPC code but it becomes explicit when you find > yourself using htonl, ntohl, and friends when building a socket > connection. There were some dual-chip jobbbies, bet-hedging. One TRS-80 could be had with a 68k board that ran CP/M-68k and as I recall the C128 had both a 65xx upgrade plus a Z80 that'd run CP/M. Dealing with big/small-endian ... that could get a bit confusing jumping back and forth between platforms. Made some boards with both a Rabbit and PIC on them and would find myself typing PIC instructions into the bunny and vice-versa sometimes :-)