Back in '85-6 I developed training for USDA (using Dan Bricklin's Demo) for a curses version of Word 5 that ran on 3B2s with SysV v2. I'd be very happy to see that ported to Linux or one of the BSDs.
In fact, I'd love to see both Word and Excel '98 on one of my workstations. IMHO things went South after that.
I had to use Windows and Office off and on with clients up to 2010 and I'm sure that Excel provided some important edge-case functionality after '98, but for me the UI went steadily to sh*t.
FWIW, Office 97 works perfectly under Wine, so there's nothing stopping you from continuing to use it, well, except maybe the lack of support for modern OOXML formats. An alternative could be Abiword and Gnumeric, both are super lightweight and retain the classic Office 97-style UI and support OOXML formats (though somewhat limited compared to LibreOffice).
https://lock.cmpxchg8b.com/wordperfect.html describes getting the UNIX release of Word Perfect running on modern Linux. There was a SCO release of Word in 1990, so plausibly that could be fudged into a similar setup?
I think it's just that 16x16 icons didn't become a thing in Windows until Windows 95, and so older apps don't supply one. Windows will then scale down the 32x32 icon instead, and it's never really done it that attractively, and I suspect the black and white nature of this icon really suffers.
Wow, it really looks like Win 10/11. Microsoft is back to the roots now.
But Word 1.0 even has 3d GUI elements. They must have had impressive computers back then.
I'm pretty sure it was an opt-in thing for individual programs, if an app didn't intentionally use it they kept the old look.
I've always thought that it's a bit of an undocumented and slightly forgotten bit of Windows 3.1, most things on the internet only show the flat style. I guess people doing retro things often only mess around a bit with the base OS, and don't really install things like Office 4.x on top of it.
The original source for WfW 1.1a was open sourced (I realize that doesn't mean there isn't copyright, but it's unlikely to be the reason that there are no pics):
The point of open source is to address and concerns around copyright. And this source is not open source by the definition most people use when they refer to “open source”. It’s still proprietary but source available.
However I will say, this source is unlikely to be useful to anyone beyond academic curiosity. So for the purposes of this project, the source license is “good enough”.
More concretely, Microsoft's license [0] gives you at most one of the four essential freedoms:
0. Freedom to run for any purpose? No.
1. Freedom to modify the program? Maybe; Microsoft can revoke it without cause at any time.
2. Freedom to distribute unmodified? No.
3. Freedom to distribute modified? No.
I'm using the free software definition [1] because it's simpler and clearer than the open source definition [2], which is usually considered mostly equivalent in practice, but as far as I can tell Microsoft's license satisfies zero of the latter's ten criteria, apart from perhaps non-discrimination and technology-neutrality. Most of the criteria of the open source definition have to do with distribution, which is not allowed by Microsoft's license.
Only if you have a lawyer department as large as theirs.
But to be fair/realistic, as an individual, I've never worried about their licenses (for myself, my friends, and family). The only time I did was when installing their software in professional/commercial contexts, that's where they do most of their enforcement as far as I know... and only when the scale is egregious.
You may use, copy, compile, and create Derivative Works of the software, and run the software and Derivative Works on simulators or hardware solely for non-commercial research, experimentation, and educational purposes. Examples of non-commercial uses are teaching, academic research, public demonstrations, and personal experimentation. “Derivative Works” means modifications to the software, in source code or object code form, made by you pursuant to this agreement.
You may copy and refer to any documentation provided as part of the software.
You may not distribute or publish the software or Derivative Works.
You may not use or test the software to provide a commercial service unless Microsoft permits you to do so under another agreement.
You may publish and present papers or articles on the results of your research, and while distribution of all or substantial portions of the software is not permitted, you may include in any such publication or presentation an excerpt of up to fifty (50) lines of code for illustration purposes.
I prompted Claude Opus 5 (I don't really have the Fable tokens to spare) to do just that (via winelib). It's gotten to the point where it opens, runs for three seconds, and abruptly exits.
edit: got it running but a ton of things are do-nothing, and the fonts are completely rubbish. Not really worth posting ATM, but I'll see if it can fix things further by running the UI tests and telling it some of the things that don't work...
(That it was able to build and run at all with just a few prompts is a HFS moment, though!)
A fair statement, but a stable ABI is only needed when you don't have the source. Seeing as most of the value proposition of Linux is that having the source is good, I find no fault in an unstable ABI.
Because the C standard is notorious for saying things like the minimum size of an `int` is 16 bits. Changes in the size of data types would create issues when calling the Windows API or saving data to disk. Heck, the program is from the very early 1990's. C wasn't standardized until 1989 and compilers were notorious for having their own interpretation of C (what we would call extensions these days). Finally, developers often optimize large programs with machine dependent code. In the early 1990's, a full featured word processor would be considered a very large program.
Even today, C is not a terribly portable language when compared to it's modern peers. It is portable when compared to assembly languages and languages contemporary to its creation.
My first guess (and probably the first guess of many of the old school nerds) was that the MS Word codename is from Opus the penguin from Berkeley Breathed's Bloom County comic strip and that seems to be correct.
I came here to say that. Though not in a bad way; the economics of what is possible but practical/impractical changes greatly with LLMs, and I suspect and hope we will be figuring out the ramifications for a while.
Agreed... I suspect we're not too far off from being able to say "port this Win16 application to Win64" if we're not there already... and most of it will be harness improvements!
In fact, I'd love to see both Word and Excel '98 on one of my workstations. IMHO things went South after that.
I had to use Windows and Office off and on with clients up to 2010 and I'm sure that Excel provided some important edge-case functionality after '98, but for me the UI went steadily to sh*t.
https://www.neowin.net/news/developer-ported-word-for-window...
https://external-content.duckduckgo.com/iu/?u=https%3A%2F%2F...
I've always thought that it's a bit of an undocumented and slightly forgotten bit of Windows 3.1, most things on the internet only show the flat style. I guess people doing retro things often only mess around a bit with the base OS, and don't really install things like Office 4.x on top of it.
Versus the port: https://cdn.neowin.com/news/images/uploaded/2026/08/17861716...
https://github.com/danielcosta/MSWORD
However I will say, this source is unlikely to be useful to anyone beyond academic curiosity. So for the purposes of this project, the source license is “good enough”.
0. Freedom to run for any purpose? No.
1. Freedom to modify the program? Maybe; Microsoft can revoke it without cause at any time.
2. Freedom to distribute unmodified? No.
3. Freedom to distribute modified? No.
I'm using the free software definition [1] because it's simpler and clearer than the open source definition [2], which is usually considered mostly equivalent in practice, but as far as I can tell Microsoft's license satisfies zero of the latter's ten criteria, apart from perhaps non-discrimination and technology-neutrality. Most of the criteria of the open source definition have to do with distribution, which is not allowed by Microsoft's license.
[0] https://computerhistory.org/blogs/microsoft-research-license...
[1] https://www.gnu.org/philosophy/free-sw.en.html
[2] https://opensource.org/osd
But to be fair/realistic, as an individual, I've never worried about their licenses (for myself, my friends, and family). The only time I did was when installing their software in professional/commercial contexts, that's where they do most of their enforcement as far as I know... and only when the scale is egregious.
https://github.com/danielcosta/MSWORD/blob/master/LICENSE
says
You may use, copy, compile, and create Derivative Works of the software, and run the software and Derivative Works on simulators or hardware solely for non-commercial research, experimentation, and educational purposes. Examples of non-commercial uses are teaching, academic research, public demonstrations, and personal experimentation. “Derivative Works” means modifications to the software, in source code or object code form, made by you pursuant to this agreement. You may copy and refer to any documentation provided as part of the software. You may not distribute or publish the software or Derivative Works. You may not use or test the software to provide a commercial service unless Microsoft permits you to do so under another agreement. You may publish and present papers or articles on the results of your research, and while distribution of all or substantial portions of the software is not permitted, you may include in any such publication or presentation an excerpt of up to fifty (50) lines of code for illustration purposes.
There is now a published GitHub of a derivative work.
edit: got it running but a ton of things are do-nothing, and the fonts are completely rubbish. Not really worth posting ATM, but I'll see if it can fix things further by running the UI tests and telling it some of the things that don't work...
(That it was able to build and run at all with just a few prompts is a HFS moment, though!)
Claude then suggested using mingw instead, which makes sense...
Even today, C is not a terribly portable language when compared to it's modern peers. It is portable when compared to assembly languages and languages contemporary to its creation.
but some don’t
agentic engineering would make this easy, especially when following the blueprint of the projects that do this translation already
Not to be confused with the Anthropic LLM model Opus - what was once an old codename is as good as new once again!
High five for anyone who still remembers the brief Bill the Gates arc in the comic or the Banana Computer parody ad of Apple starring Gene Simmons (https://computerhistory.org/blog/the-banana-junior-6000-comp... , about halfway down the page).
Also, a direct cognate with "operate", "operation", etc.
(Also some Norwegians and a Web browser?)