Twitter | Search | |
Jonathan Blow
Game designer of Braid and The Witness. Partner in IndieFund.
1,190
Tweets
98
Following
119,590
Followers
Tweets
Jonathan Blow 5h
Replying to @Bigcheesegs
Nope. This is the whole problem with this mindset. It *does* matter, because the actual job I am trying to do is put a specific program onto a specific computer. It's not that toxic in this case, but the whole UB cluster-f is super toxic.
Reply Retweet Like
Jonathan Blow 6h
It's great that people are enthusiastic, but I do not really like this, not least because the information is often not presented with the right emphasis to communicate the real ideas. But that is just the cost of doing things the way I am doing them, I guess.
Reply Retweet Like
Jonathan Blow 8h
An engine failure is not an exception, it is the expected case. You should handle it explicitly in code, just like, say, the disk being full when you try to write a file. Anything else is just asking for bugs and crashes.
Reply Retweet Like
Jonathan Blow 12h
Yep.
Reply Retweet Like
Jonathan Blow 12h
For what it’s worth, I don’t think LLVM will ever be objectively fast; its API seems slow by design. It can get faster than it is, for sure, but, ehh.
Reply Retweet Like
Jonathan Blow 12h
That’s good to hear, at least.
Reply Retweet Like
Jonathan Blow 12h
If your system does not run in a reasonably consistent amount of time, that is your first problem and you should fix it.
Reply Retweet Like
Jonathan Blow 12h
Replying to @nathan_k
I think caring a lot about error messages is a very good idea.
Reply Retweet Like
Jonathan Blow 12h
Nope. When I talked to whoever was running the compiler group some years back, his reply was “It shouldn’t be slower; we are not doing anything that should make it slower.” I was like, “Do you measure the speed?” and he was like, “No.”
Reply Retweet Like
Jonathan Blow 12h
the compilers, etc.
Reply Retweet Like
Jonathan Blow 12h
The problem is, we don’t seem to have that most of the time. e.g. Visual Studio is very much slower at compiling C++ than it used to be, (actually also gcc??), even for code without templates, and nobody seems to notice that anything is wrong, especially not the authors of
Reply Retweet Like
Jonathan Blow 12h
I think that’s fine, as long as adult supervision exists that understands this is a short-term fix and not a permanent dookie in the pile.
Reply Retweet Like
Jonathan Blow 12h
I am not a fan of a lot of the stuff that modern compiler authors are doing ... but ThreadSanitizer is very valuable.
Reply Retweet Like
Jonathan Blow 13h
Adding caching increases the complexity of the system. It is not a good way to solve problems that should not be there in the first place. The right thing to do is to fix the actual problem.
Reply Retweet Like
Jonathan Blow retweeted
Donald Mitchell 15h
Came across some of my old CCITT/OSI books. I think the theory of the OSI stack is still taught in the classroom, but I suspect a lot of the lore will be forgotten.
Reply Retweet Like
Jonathan Blow 15h
Dude I have been using Linux since the early 1990s. I am not talking about one instance of one problem. If that's the problem, show me the one that works really great.
Reply Retweet Like
Jonathan Blow 16h
What about any of this is supposed to counter my point?
Reply Retweet Like
Jonathan Blow 16h
Replying to @dotstdy
You could, you know, respond to the actual point, with something of substance.
Reply Retweet Like
Jonathan Blow 20h
Replying to @scottjmaddox @tblodt
Yes but that is not very easy.
Reply Retweet Like
Jonathan Blow 21h
Replying to @sx58941646 @niinegames
Every single version of Visual Studio after vc6 has had tons and tons of bugs. These days the bug count increases over time.
Reply Retweet Like