+
+
+
+
+
+
+
+
+
+
+
+ 
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 
+
+
+
+
+
+
+
+
+
+
+
+
+
+ I hope we don't spend the entire time discussing the definition of simplicity 
but I do think it's probably good to clear up yeah.
+
+
+
+
+
+
+ but I do think it's probably good to clear up yeah.
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ Right---that alone is a rathole that could last hours 
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 
+
+
+
+
+
+
+
+
+
+
+
+
+
+ I was thinking about this definition of simplicity in anticipation of this conversation, and I do think it makes sense to outline some various ideas of what simplicity means to us. But I am going to propose we explore the way various kinds of simplicity relate to performance. That'll give us more to talk about and we don't have to agree on a single definition to get to good stuff.
+
+
+
+
+
+
+
+
+
+ 
+ 1
+
+
+ 1
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ To me I just work backwards from the reality that I want good software, and a lot of people who know what they're talking about agree that something called simple is relevant, and I arrive at something like this. Something is simple when it is easy to work with, easy to implement, easy to use, easy to understand, etc, especially when it also solves a problem that can seem hard without this "something".
+
+
+
+
+
+
+
+
+
+
+ 1
+
+
+ 1
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ when you want to do X you do X, not X and several other things as well
+
+
+
+
+
+
+
+
+
+
+ 
+ 2
+
+
+ 2
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ This is why we have @rxi here, because one of the things I've noticed is that his code is both what I would consider very simple, and probably as a result, very small (compared to what others, including myself, would probably come up with)
+
+
+
+
+
+
+
+ easy to implement, because fewer codepaths => less implementation work
+
+
+
+
+ 1
+
+
+ 1
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+The analogy to orthogonality makes me more uncomfortable. Because that sounds to me like the mental framework goes: I sit down and identify N effects I need to achieve and therefore write the N simplest codepaths I can think of. But I think that's just the beginning. Overtime it might turn out that fewer codepaths actually could do it all.
+
+
+
+
+
+
+
+
+
+
+ As a concrete example of this, the command list design of microui simplifies drawing down to just a few small primitives, where otherwise you might have to implement custom drawing functions for buttons, windows, scrollbars, etc.
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ @rxi I'm curious to what end you have "simplicity" in mind as you're designing something (however you define it)
+
+
+
+
+
+
+
+
+
+
+ 1
+
+
+ 1
+
+
+
+
+
+
+
+
+
+
+ 
+
+
+
+
+
+
+
+
+
+
+
+
+
+ I think i am alright with that definition of simplicity, i think much of what i would consider simple would be to have the least amount of accidental complexity as possible for the problem (which @gingerBill often mentions). I think few codepaths would be a good measure of that
+
+
+
+
+
+
+
+
+
+
+ 2
+
+
+ 2
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ I'm guessing it's not as heavy on the linear algebra as Ryan's thought process, but I could be wrong
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ And for whatever reason, the linear algebra is really intuitive for me when trying to do that
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ So I think I first realized all this a while back in a specific 4coder problem, thought it might be interesting to break down that problem and show how I made performance worse by being complicated.
+
+
+
+
+
+
+
+
+
+
+ 1
+
+
+ 1
+
+ 
+ 1
+
+
+ 1
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ Alright, one minute to sketch it out...
+
+
+
+
+
+
+
+
+
+ 
+ 4
+
+
+ 4
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 
+
+
+
+
+
+
+
+
+
+
+
+
+
+ hi
+
+
+
+
+
+
+
+
+
+ 
+ 4
+
+
+ 4
+
+ 
+ 1
+
+
+ 1
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ So in what sense would you consider your new design "simpler" than your previous one? It does have more layers after all
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ I realize I'm deviating from the topic of performance here 
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ Well Allen made a really important point here that I don't want to miss, which is that something that is easier to maintain implies that it requires less time and less effort to maintain, which implies a larger number of iterations and a more manageable set of tasks to layer optimizations into.
+
+
+
+
+
+
+
+
+
+ 
+ 1
+
+
+ 1
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ This is it. If I can't get anything else across from that story, I would hope this still makes it. Your capacity for handling ever increasing difficult problems is a finite resource.
+
+
+
+
+
+
+
+
+
+
+ 4
+
+
+ 4
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+on that note, I suppose we could also explore our experience with systems that are hard to get good performance out of
+
+
+
+
+
+
+
+
+
+
+ Trying to brainstorm about this. Frankly I don't have a whole lot of experience when it comes to this kind of thing, mostly because most of the programming problems I have dealt with in life have been invented by me, so I can conveniently change the constraints, or avoid nasty performance problems more easily.
+
+But let's say you're working with, you know, some file format designed by some people somewhere... hypothetically speaking... in some theoretical scenario. You're now locked into those constraints if you hope to be working with files stored in that format.
+
+The characteristics I think of here have to do with "distance", between some pattern of data extraction, and how that data is stored. Maybe what I am getting at is how much work goes into decompression.
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ Well so I could talk directly about laughably slow things at work if we need to
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Well so I could talk directly about laughably slow things at work if we need to
+
+
+
+
+
+
+
+
+
+
+
+
+ I am interested in this too
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ I'll give the quick outline, but here's the additional reading if you really want to understand: https://fgiesen.wordpress.com/2018/03/05/a-whirlwind-introduction-to-dataflow-graphs/
+
+
+
+
+
+
+
+
+
+
+
+ 1
+
+
+ 1
+
+
+
+
+
+
+
+
+
+
+ So it is at least another example that Complexity + RunningTime is not a conserved quantity. There are win-wins.
+
+
+
+
+
+
+
+
+
+
+
+ 1
+
+
+ 1
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+I think I am getting lost in the mines of moria, as it were
+
+
+
+
+
+
+
+
+
+
+ Yeah maybe this is too in the weeds :P
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Yeah I think this is probably a bit out there. Hard to talk concretely about DWARF without first spending a really long time on the introduction to the problems.
+
+
+
+
+
+
+
+
+
+
+ Yeah, probably not the best example, I just brought it up because that is the most notable example I've had of "systems that seem complicated that have performance problems", and I was trying to brainstorm. All of my personal programming work has not really dealt with this nearly as much.
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+I can tell you though that for some of what I've had to deal with, it's very related to simplicity
+
+
+
+
+
+
+
+
+
+
+ I am interested to hear more on this
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ So that's kind of my whole thesis on the issue. I don't think simplicity and performance are directly related much at all. But I think simplicity is critical if you want to actually have any hope of addressing performance problems.
+
+
+
+
+
+
+
+
+
+
+
+ 1
+
+
+ 1
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ Alright looks like we've about exhausted #fishbowl talk, maybe time to digest the conversation! Thanks everyone for participating, was really interesting.
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ [End of Fishbowl]
+The beginning of this conversation has been pinned, so if you'd like to read it from the start, go there! See you next time.
+
+
+
+
+
+
+
+
+
+
+ 1
+
+
+ 1
+