Comment by axelriet
7 years ago
It’s my personal opinion and it’s based on my lecture of ReactOS code at the time I downloaded it (circa when I wrote my reply on Quora). I think anyone who can read C code can reach similar conclusions just by eyeballing the code in ReactOS and the code from the leaked Research Kernel, that can be found for example on GitHub. As a matter fact, someone on this thread did just that and it did not take that person very long (just minutes) to find some of the similarities that I, for one, qualify as troubling. It should not be a problem for anyone to repeat the experiment, all code is available for download. So why not go ahead and see that code with your own eyes? I think it’s highly improbable that a reimplementation of that magnitude came so close from the original in so many respects. I’ve been conducting interviews for years and screened some of the best and most promising future engineers coming from the best universities and I always ask the same coding questions. I was given many decent and correct answers, but never the actual implementation was similar to the level I can observe between ReactOS and the leaked Microsoft code. In fact every candidate wrote very different answers to the very same question, some even had very different approaches to the resolution of the same problems. I cannot explain how a team of people on a hyper-complex open source project can write nearly the same code - the source code looks very similar down to peculiar formatting idiosyncrasies - as a separate team of developers of the original closed-source project, just from observing the external behavior and/or reading the documentation, if available at all. I think this is impossible.
>and it’s based on my lecture of ReactOS code at the time I downloaded it
It would be interesting to see if this statement comes back to bite you in court if the opposite accusation is ever made. That Windows kernel developers have been stealing from GPL code :)
Man, I thought the exact same thing. I must have stolen it from you. Sorry!
In all seriousness is this normal? I write and work on kernel drivers for the company I work for and have always gone out of my way to not view or interact with any code that remotely relates to the work I do (out side of what I coded). Our competitor even released their version of a block device snapshot driver and while I would be tempted to see how they did things I have avoided it every step of the way.
So, if a Microsoft kernel developer doesn't follow the practice of not reading kernel code with an incompatible license, does that mean we can assume they don't follow it generally? Can we assume they take code from the Linux kernel, violating the GPL?
6 replies →
I presume you quit working on Windows after you read the ReactOS source code, offered to you under the GPL license only? Or is Microsoft planning to release it under a compatible license?
Plot twist: Microsoft employee goes out of his way to force Microsoft to open-source Windows
Your interview analogy is a bad one. Because they are not trying to make their solutions as similar as possible to each other.
If you had interviewee A write some piece of code, and then said to interviewee B "I want you to solve this problem as close as possible as the person before you did it, here is their executable", you'd be quite surprised what they could do. Especially if interviewee B is Alex Ionescu.
I don't suppose you ever compared parts of the tree that weren't present in the various sort-of-public NT source leaks or research kernel, to see if idiosyncrasies there matched as well?
> It’s my personal opinion
Others in the thread have pointed out several verifiably false fact claims in it, so it's both derogatory and at least in part factually false, not a mere matter of subjective opinion.