I honestly haven't read too many justifications (that's partly what spawned this question in my mind). I remember a time when there was ctags in Vim for this kind of stuff. The LSP paradigm, for me, feels easier. But I am curious if I am missing something?
-
-
Replying to @tahini @Jonathan_Blow
Put another way, when I think 'bloated and slow', I typically think of an actual IDE.
1 reply 0 retweets 0 likes -
Replying to @tahini
Okay, so, how is it that an actual IDE gets 'bloated and slow'?
1 reply 0 retweets 1 like -
Replying to @Jonathan_Blow
Continual expansion of features, driven by monetization. Complexity, basically.
1 reply 0 retweets 0 likes -
Replying to @tahini
Really? Does Visual Studio of today have a lot more features than Visual Studio did in 2005?
1 reply 0 retweets 2 likes -
Replying to @Jonathan_Blow @tahini
Like enough features to compensate for Moore's Law? Like 100x as many features?
1 reply 0 retweets 3 likes -
Replying to @Jonathan_Blow
No, not 100x as many features. Not enough to compensate for Moore’s Law.
2 replies 0 retweets 0 likes -
Replying to @tahini @Jonathan_Blow
I get that programmers today are spoiled by processing power and that has made them inferior programmers when it comes to perf — I’m just trying to get at how fact that is particularly highlighted with LSP
1 reply 0 retweets 0 likes -
Replying to @tahini @Jonathan_Blow
I think Jon was just hoping most programmers would know the rough cost delta between a direct call using a shared memory binary AST, and a socket-mediated RPC that requires serializing the data to and from UTF-8 both ways. LSP is just sad proof that they definitely don't.
1 reply 0 retweets 6 likes -
Replying to @cmuratori @Jonathan_Blow
Was LSP implemented in manner that is performant? No, it doesn’t look like it. But it exists, and I think it’s popularity has to do with the lack of viable alternatives. If implicit suggestion here is to just use an IDE or ctags ... well many people don’t want this things.
1 reply 0 retweets 0 likes
But this has nothing to do with IDEs. You can trivially design a similar AST interchange protocol that is in-memory, binary, and does not require sockets. The sadness (presumably) comes from the fact that nobody seemed to think that was an important consideration.
-
-
Replying to @cmuratori @Jonathan_Blow
Ok this is what I was looking for. So there was a recognized need as an industry to come together to make some sort of AST standard like this, and the chosen implementation (out of presumably others) was bad.
2 replies 0 retweets 0 likes -
Does socket-based approach come with any advantages at all over the alternatives?
1 reply 0 retweets 0 likes - Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.