> I think they crapped over those tools, because these tools got the abstraction wrong.
If you identify a problem in a useful tool, the correct response isn't "throw it into the garbage and never make a tool anything like it ever again!" The solution is to research and iterate and try to fix the problem. The Wright Flyer didn't have much range. The Wright's solution wasn't "let's give up on creating airplanes forever!"
We'll never know if Access could be made better, because development on it stopped.
And in any case, for the point I'm making, it doesn't matter if Access is a perfect, or even a good tool. The point I'm making is that he developers of Access had a clear vision of how to make software development easier for the average person that now no longer exists. Maybe they weren't good at achieving that vision, but they had it. Modern software developers, by-and-large, do not. (HyperCard's a far better example, since it had the same vision and also more successfully executed on it.)
> The problem git solves is just non-trivial, so git itself has some learning curve.
Microsoft Word solves the same problem Git does with its "Track Changes" feature and is far easier to use. I'm not saying that "Track Changes" is sufficient for software development purposes, but it certainly points to a few obvious ideas that could make source control software far better.
But of course Git has none of that. It doesn't even try. It's simply bad software.
EDIT: BTW, the problem with Git isn't that it's too complex, the problem is that it's far too simple. It implements a few abstractions, then forces the user to deal with them even though those abstractions don't resemble human thought processes even slightly.
And for the record, Subversion solved the exact same problem and was far more usable. It also had more features. And was much easier to create GUIs for. Basically, it was better than Git in every measure that matters to me. And it was bad software. But it was way better than Git.
If you identify a problem in a useful tool, the correct response isn't "throw it into the garbage and never make a tool anything like it ever again!" The solution is to research and iterate and try to fix the problem. The Wright Flyer didn't have much range. The Wright's solution wasn't "let's give up on creating airplanes forever!"
We'll never know if Access could be made better, because development on it stopped.
And in any case, for the point I'm making, it doesn't matter if Access is a perfect, or even a good tool. The point I'm making is that he developers of Access had a clear vision of how to make software development easier for the average person that now no longer exists. Maybe they weren't good at achieving that vision, but they had it. Modern software developers, by-and-large, do not. (HyperCard's a far better example, since it had the same vision and also more successfully executed on it.)
> The problem git solves is just non-trivial, so git itself has some learning curve.
Microsoft Word solves the same problem Git does with its "Track Changes" feature and is far easier to use. I'm not saying that "Track Changes" is sufficient for software development purposes, but it certainly points to a few obvious ideas that could make source control software far better.
But of course Git has none of that. It doesn't even try. It's simply bad software.
EDIT: BTW, the problem with Git isn't that it's too complex, the problem is that it's far too simple. It implements a few abstractions, then forces the user to deal with them even though those abstractions don't resemble human thought processes even slightly.
And for the record, Subversion solved the exact same problem and was far more usable. It also had more features. And was much easier to create GUIs for. Basically, it was better than Git in every measure that matters to me. And it was bad software. But it was way better than Git.