Is Code Reuse Bad?

This is a conversation between Huiqi and I about the notion I have heard recently from multiple very intelligent people that software reuse is bad, which flies in the face of most current opinion. Here are my sources: https://www.youtube.com/watch?v=7YpFGkG-u1w https://www.youtube.com/watch?v=Ge3aKEmZcqY https://www.forth.com/resources/forth-programming-language/ https://www.informit.com/articles/article.aspx?p=1193856

Transcript

The following transcript was automatically generated by an algorithm.

00:01 : so design patterns and solid are
00:05 : concepts we've been talking about
00:07 : um but very recently
00:09 : um i've been
00:11 : reading and watching
00:14 : some materials
00:17 : that go against them
00:20 : there's a video i watched that blatantly
00:23 : blames solid
00:25 : for
00:26 : why software is bad nowadays
00:30 : and i want to
00:32 : dig into that perspective
00:34 : because i i do think there is something
00:36 : there
00:38 : one of them is
00:40 : casey
00:42 : mor
00:43 : moratorium
00:45 : his channel is called molly rocket he's
00:48 : a game developer
00:50 : and he's like extremely smart
00:52 : okay um
00:54 : he's a has a big following
00:57 : so where does that code come from
00:59 : this is the this video i was talking
01:02 : about he also has a separate one that i
01:04 : liked
01:06 : called
01:07 : simple code
01:09 : high
01:10 : performance i really like this video
01:13 : it's a three hour lecture
01:15 : okay uh but i really really like that
01:18 : but in in
01:20 : has several things in it that i liked
01:22 : but one of them
01:24 : is a point where he says
01:28 : basically he says
01:29 : code reuse is
01:32 : bad more or less
01:35 : this concept of uh could reuse being a
01:38 : bad thing uh
01:41 : i want to say i haven't
01:43 : really heard it before but recently i
01:46 : have heard it like three or four times
01:49 : from different people
01:51 : yeah so this
01:53 : this part where he says
01:56 : um
01:57 : that's a lot of code the meta reason
02:00 : so he's talking about a piece of code
02:02 : that is slow and he rewrote it to be
02:05 : fast
02:06 : like a lot faster
02:08 : like okay like more than 10 times faster
02:11 : maybe maybe 30 times faster or even more
02:14 : um and he said
02:16 : why is the code slow the meta reason
02:20 : when the why code is slows because
02:22 : because people write way too much code
02:25 : they do this in two ways number one is
02:28 : they reuse a lot of code reusing code is
02:31 : something that people have been taught
02:32 : to do
02:34 : and there are
02:35 : reasons why you may want to reuse code
02:37 : but the problem with it is when you
02:39 : reuse code you usually don't think about
02:41 : all the stuff that code is doing
02:44 : that you didn't actually need needed to
02:47 : do basically using like reusing code
02:49 : that's doing more
02:52 : than what you need
02:54 : for that use case that for that
02:57 : situation
02:58 : that you have
03:00 : um
03:01 : and
03:02 : yeah because when you use a library like
03:05 : the say you use a complex library and
03:08 : the library is trying to resolve all
03:10 : these issues for everybody because it's
03:12 : a library but you only resolve yeah the
03:15 : library has to check for error
03:18 : conditions right that they're like oh is
03:21 : your input even correct or like oh maybe
03:24 : the library is in
03:27 : you put the library in a different mode
03:29 : or there's like if it's a css like
03:33 : you're in the flex mode or you're in
03:36 : grid mode or you're in the
03:38 : normal
03:40 : css mode and depends on which mode it is
03:42 : you can do this or you can't do this um
03:46 : and all those if statements
03:48 : are not things you would need if all you
03:51 : ever have is
03:52 : flex mode for example
03:55 : um
03:59 : so
04:00 : so that's what he's saying like like
04:02 : we're doing code reuse too much and it
04:06 : has these drawbacks
04:08 : yeah i mean
04:09 : um
04:11 : i i think yeah i mean i kind of agree
04:14 : with what he's saying that maybe we
04:17 : reuse could too much
04:19 : and another case that is not um that is
04:23 : different from like performance is uh
04:26 : you remember i went to the architecture
04:28 : like conference at some point uh with
04:31 : jason
04:32 : new york and that conference is all
04:35 : about like microservice and stuff like
04:37 : that so basically the idea is
04:39 : microservice at the end is all the
04:42 : service they don't depend on each other
04:45 : they don't have like
04:47 : shared components so they're like
04:49 : completely independent
04:51 : so that's the first time i hear like oh
04:54 : you should not reuse code you should
04:55 : just copy the code yeah yeah
04:58 : yeah
04:59 : just because of the independency yeah i
05:02 : remember that because
05:04 : because the dependency like i'm
05:07 : depending on a piece of code that does
05:09 : something for
05:10 : me
05:11 : but now i have to do something for it
05:14 : um so so two more two more sources that
05:18 : i want to put here
05:21 : so one one is so i was really digging
05:23 : into this fourth programming language
05:26 : which again we will uh
05:28 : um
05:30 : i'll put out more compilers challenges
05:33 : okay and it will cover this language
05:36 : because i i am i'm really in love with
05:39 : this language
05:40 : the the the the original inventor of
05:43 : fourth his name is chuck moore or
05:46 : charles moore
05:48 : and there's a whole story about him but
05:50 : uh it talks about his three principles
05:54 : that he uses
05:55 : for coding
05:57 : okay um
05:59 : first of all to more forth was a
06:01 : personal response to his frustration
06:04 : with existing software tools which he
06:07 : viewed as a sort of tower of babel
06:10 : and
06:11 : the word babel i think he used
06:14 : interestingly that's actually a tool we
06:16 : use
06:17 : in the javascript community but i think
06:20 : the word babel he uses to mean
06:23 : um
06:24 : there's a lot of languages and you have
06:26 : to translate
06:29 : like to and from a lot of levels of
06:32 : translation
06:33 : so and and babel is
06:36 : okay according to the story a united
06:39 : human race
06:41 : in generations
06:42 : following the great flood speaking a
06:45 : single language
06:47 : and migrating eastward comes to the
06:51 : okay
06:53 : okay i have to read this but but um but
06:55 : maybe not right now but uh but i guess
06:58 : yeah that the frustration is there's a
07:01 : lot of languages
07:03 : like like the compiler is written in a
07:05 : different language from
07:07 : the language that's actually
07:09 : you're using to write your application
07:13 : and then the compiler has to translate
07:15 : your language
07:17 : into
07:18 : the assembly language
07:20 : so there's all these layers of
07:22 : translation going on and
07:25 : you know i want to teach people
07:27 : compilers
07:28 : but it's difficult
07:30 : precisely because there's all these
07:32 : languages that they're going to have to
07:34 : learn
07:37 : and learning one language is difficult
07:40 : enough right you have to learn
07:43 : several languages
07:44 : above all his guiding principle which he
07:48 : calls the
07:49 : basic principle is
07:51 : keep it simple
07:53 : um
07:55 : okay
07:57 : that makes sense
07:59 : that makes sense okay
08:01 : so number of capability you add to the
08:03 : program increases the complexity of the
08:06 : program increases exponentially
08:09 : you know it can easily go get out of
08:11 : hand you can avoid this if you apply the
08:14 : basic principle
08:16 : um
08:17 : i'll say in practice you know many it
08:20 : just takes more effort to
08:23 : always come up with the simple solution
08:27 : so if you're always rushing software out
08:29 : the door
08:30 : um
08:32 : you're not gonna take that extra effort
08:34 : and time
08:35 : to to do it the way that's the simplest
08:38 : and that that's what i was sort of
08:40 : talking about
08:41 : earlier when i said i just go out for a
08:44 : walk and i come back with some ideas
08:48 : about how to make it simpler right
08:50 : but that requires
08:52 : more
08:53 : patience and time
08:55 : well i guess the other thing is the
08:58 : requirement just become more and more
09:00 : complicated if you think about like the
09:02 : software people
09:04 : had like say 30 years ago it's not that
09:07 : complicated like so the front end is
09:09 : just rendering some text and then you
09:11 : can
09:12 : render the text and be done with it but
09:14 : now it just gets crazy like you need all
09:17 : these kind of stuff like interactions
09:20 : all this formatting all this css
09:22 : everything just gets more and more
09:23 : complicated
09:25 : but but are those are all those
09:28 : layers necessary to solve the problem
09:32 : or are they just
09:34 : are they just complications that were
09:37 : put there
09:39 : accidentally are they accidental
09:42 : complexity
09:43 : or are they inherent complexities
09:46 : i think that's debatable
09:48 : okay like we don't
09:50 : we don't have to use css
09:53 : to lay out our stuff
09:56 : right
09:57 : that's true if if
09:59 : it's been said that like css is not
10:02 : necessarily the greatest
10:04 : um
10:05 : well
10:06 : i'm not going to get into that but
10:08 : should i give if you could redesign css
10:11 : from scratch or don't even think about
10:14 : css it's not it doesn't have to look
10:17 : anything like css at all
10:19 : you're going to redesign a way to lay
10:22 : out your things
10:24 : how would you do it and could you do
10:26 : could you come up with a simpler scheme
10:28 : than that
10:30 : you probably can
10:32 : um
10:34 : uh okay his second principle is
10:37 : do not speculate
10:40 : yeah i i say that as well uh but
10:43 : yeah that's the you ain't gonna need it
10:46 : principle
10:47 : of tdd
10:49 : like don't don't
10:51 : yeah don't build anything that you might
10:53 : not need
10:55 : um
10:56 : and then his third principle is do it
10:58 : yourself
11:00 : his he's basically going against the
11:03 : idea of code reuse in a way
11:06 : he tends to
11:08 : like rewrite his own stuff all the time
11:11 : like
11:13 : before you can write your own
11:14 : subroutines you know have to know how
11:16 : this means to be practical that you have
11:19 : written that before
11:20 : which makes it difficult to get started
11:23 : but give it a try after writing the same
11:25 : subroutine a dozen times
11:27 : on as many computers and languages
11:30 : you'll be pretty good at it
11:32 : um
11:34 : yeah so it keeps rewriting the same
11:36 : stuff over and over again each time with
11:39 : a variation
11:41 : and uh like well it's obvious that you
11:44 : would get really good at it if you kept
11:46 : doing that for years
11:48 : um
11:51 : i think that's interesting
11:54 : and then another one is this
11:56 : this uh
11:57 : [Music]
11:58 : interview with
12:00 : uh donald knuth sorry uh and he's the
12:04 : author of this very famous series called
12:07 : the art of computer programming
12:09 : so
12:10 : andrew asks
12:12 : donald knuth
12:14 : uh
12:15 : on you on your website you mentioned
12:17 : that
12:19 : blah blah blah
12:21 : in the segment 93 called advice to young
12:24 : people you advise that people shouldn't
12:26 : do something simply because it's trendy
12:30 : um as
12:32 : as we all know too well software
12:34 : development is as subjective
12:36 : as subject to fads
12:39 : as any other discipline
12:42 : can you give some examples that are
12:44 : currently invoked
12:46 : which developers shouldn't adopt simply
12:49 : because they are currently popular or
12:52 : because that's the way they're currently
12:53 : done blah blah and he says
12:56 : that question is almost contradictory
12:58 : because i'm basically advising against
13:02 : advising young people to listen to
13:04 : themselves rather than to others and i'm
13:07 : one of those others so if you listen to
13:10 : me then you wouldn't be listening to
13:12 : yourself
13:13 : but but here later on he says
13:18 : i must also confess to a strong bias
13:22 : against the fashion for reusable code
13:26 : to me re-editable code is much much
13:29 : better
13:30 : than an untouchable black box
13:33 : or toolkit
13:35 : i could go on and on about this
13:38 : if you're totally convinced that
13:40 : reusable code is wonderful i probably
13:43 : won't be able to sway you anyway but
13:46 : you'll never convince me
13:49 : that
13:49 : reusable code isn't mostly a menace
13:55 : okay
13:57 : mostly a menace you're saying reusable
13:59 : code is mostly a menace that's a very
14:02 : strong word
14:04 : um
14:06 : so yeah i i i think that's
14:10 : i think that's very interesting like
14:12 : we've all been taught like
14:14 : reusable code is
14:17 : good something you should strive for
14:19 : and uh
14:21 : i i wrote like a story about the three
14:23 : little big story about it
14:26 : right
14:27 : but we have these very
14:29 : smart
14:31 : and experienced individuals that
14:35 : that really go against the idea of
14:37 : reusable code
14:39 : so i think there's something there
14:42 : well i would say if i have unlimited
14:45 : time and no like deadlines of like
14:48 : pushing new software or like oh you have
14:51 : to
14:52 : deliver this functionality by this time
14:54 : if that is not a thing i agree with him
14:58 : i want to like just pull like other
15:00 : people's library and do it probably will
15:03 : try to do it myself and
15:06 : because i will learn it if i do it
15:07 : myself and i can edit if i need to and i
15:10 : can add more stuff but i think most of
15:13 : the times i have to use other people's
15:16 : thing it's just because
15:18 : i don't have enough time to reinvent the
15:21 : will in the work setting
15:23 : i think that's
15:25 : what to me is a
15:27 : an issue
15:29 : like you you feel like you have to
15:31 : because there are these things that are
15:34 : available
15:35 : out there
15:36 : that you can use
15:38 : and i think there's also this
15:41 : peer pressure of
15:43 : your co-workers
15:45 : know that there are these things that
15:47 : you can use
15:49 : why why wouldn't you use them
15:51 : that kind of thing
15:53 : um
15:54 : or or they're just these uh very
15:58 : popular libraries
16:00 : that maybe the company really trusts
16:03 : yeah so to go against the grain like
16:06 : that
16:07 : uh might not be a popular choice to make
16:11 : i think those things are there um
16:14 : i i think there might be situations
16:16 : where
16:18 : it would
16:19 : actually
16:20 : not not every situation obviously but
16:22 : there are there can be situations where
16:25 : it would actually be more efficient to
16:28 : just do that thing yourself instead of
16:30 : use a library that's available because
16:32 : maybe the library is not that good and
16:35 : doesn't match
16:36 : your needs
16:38 : perfectly
16:39 : yeah i have encountered that situation
16:42 : before
16:43 : as well so which is oh there is actually
16:46 : a library that does the thing but
16:48 : actually uh the thing that i created is
16:53 : simpler and because we created in-house
16:56 : so you can like add functionality if you
16:57 : want to or you can change the syntax
16:59 : change the interface if you need to
17:01 : but the library you just can't do it so
17:04 : at the end we go with the in-house
17:06 : solution but it's case by case i don't
17:09 : think it's just like always this or that
17:11 : yeah