Doing Windows, Part 3: A Pair of Strike-Outs | Hacker News Reader

Doing Windows, Part 3: A Pair of Strike-Outs

filfre.net

64 pointsdoppp9 comments

I love the roast they did to apologize for Windows being vaporware for so long.

> Gates said that Ballmer had wanted to cut features: “He came up with this idea that we could rename this thing Microsoft Window; we would have shipped that a long time ago.”

The interesting question remains: why did they keep going? Windows 1 failed horribly (just like similar products), why make a second version? And secondly, why make it compatible with windows 1 if nobody was using it anyways. This is usually the perfect moment to fix old mistakes, change the architecture, etc. It was all there in windows 1... GlobalAlloc, SendMessage, etc.

You can take a windows 1 binary, patch the PE header to say it's for windows 2, use Borland resource editor to recompile any embedded resources, and then it will run on 32bit windows 10.

This is speculation, but anybody who was using windows 1 was either a fanboy, or used a must-have windows 1 application. Pissing of those people is a bad way to introduce your next version.

They did make a few changes, such as in window hooks. Compare http://www.os2museum.com/files/docs/win10sdk/windows-1.03-sd... and http://www.os2museum.com/files/docs/win20sdk/windows-2.0-sdk... for example. Actually, this makes me wonder if they never bothered with Windows 2.x and went directly to OS/2 Presentation Manager in 1987.

What architecture changes would you have wanted for Windows 2.0, given the constraint of running on an original 8086? (IIRC, that constraint wasn't lifted until Windows 3.0.)

This was more of a general comment. I mean sure, if you couldn't have designed it any better given the technical constraints this makes sense, but just from personal experience there is always something you wish you'd have done differently in retrospect. This is why I can only assume there must have been at least one critical customer they didn't want to piss off, or it was somewhat of a proof of dedication to backwards compatibility, as in "see, even if nobody is using our product we still make the next version compatible".

This is a fantastic, in-depth series (I can't wait for more!) It really shows how much success is based on luck and timing. Be sure to read the first two parts, if you haven't already: https://www.filfre.net/2018/06/doing-windows-part-1-ms-dos-a...

A very enjoyable read, but it's a shame there was no mention of Digital Research GEM, which did ship in 1985. I remember trying it on various Amstrad PC machines in the UK and being very impressed. I never saw an Apple Macintosh at that time, even for sale in a shop, probably due to its price. A few years later GEM starting shipping with Atari ST machines. I presume the fate of GEM was ultimately tied to that of DR.

He has already done an article on Atari ST, and GEM is included there.

https://www.filfre.net/2015/04/the-68000-wars-part-2-jack-is...

He is basically working through the years, touch upon the major topics of each one as he goes. I suspect that he is building some background on Windows 3.x, the first really successful version of Windows, with these articles.

Reply on news.ycombinator.com