Clever vs Savvy

This is a presentation I gave to a company where I work as a software mentor. The talk is about software maintenance and code re-use. It is entitled: Clever vs Savvy.

Transcript

The following transcript was automatically generated by an algorithm.

00:02 : so the title of this talk is clever
00:06 : versus savvy I was sitting down with
00:08 : Colin and he asked me like so what what
00:14 : do you mean by clever and savvy are they
00:17 : like are they like software tools that
00:21 : you download no no I mean the actual
00:27 : words and he was like well clever and
00:31 : savvy feels like pretty similar to me
00:34 : it's like no they should not feel
00:38 : similar or at least that's what doesn't
00:41 : not my intention so so then I googled up
00:44 : their definitions and let's try to get
00:50 : into the root of these two words so
00:57 : cleverness
00:58 : it means quick to understand that's
01:01 : actually not exactly what I'm trying to
01:03 : get at I think this is really what I'm
01:06 : getting at clever what I mean by clever
01:11 : in this talk is showing intelligence or
01:15 : skill ingenious like you're you're yours
01:20 : you're showing off your the ability of
01:24 : your mind you're doing something super
01:27 : difficult it's sort of like the
01:30 : equivalent of lifting a lot of weight at
01:35 : the gym to show off your physical
01:39 : prowess it could be used to mean like
01:42 : you're showing off your mental prowess
01:44 : perhaps whereas savvy I'm using savvy to
01:50 : mean that shrewdness and
01:57 : practical that the keyword being
01:58 : practical practical and common sense
02:04 : common sense because I'm actually to be
02:08 : more specific I'm actually going to use
02:09 : savvy to mean knowledge about the
02:14 : business domain right business knowledge
02:16 : and I'm actually using these adjectives
02:20 : for code artifacts for code that we
02:24 : write if we were to generalize it beyond
02:28 : the domain of software it could just be
02:32 : applied to written artifacts articles or
02:37 : descriptions if we're talking about UX
02:39 : it could be the description and the
02:42 : diagrams that you're making for the
02:45 : purpose of UX coming it's communicating
02:47 : you its ideas so so if a piece of code
02:54 : is savvy what I mean is that piece of
02:57 : code has a lot of business knowledge
03:00 : embedded in the code and a good way to
03:03 : tell that is you are reading the code
03:06 : and the vocabulary of the code consists
03:10 : of a lot of terms from the business
03:14 : domain so if we're talking about a
03:17 : project management system then you see
03:21 : words like project teams schedules
03:27 : deadlines status words like that which
03:30 : are all vocabularies of the business
03:33 : domain the opposite of savvy so if this
03:36 : were a scale this true this to the left
03:40 : is the most savvy to the right would be
03:44 : the most ignorant a piece of code that
03:47 : is ignorant would be code whose
03:51 : vocabulary is devoid of any business
03:54 : knowledge devoid of any terms that's
03:58 : related to any business related I
04:01 : yeah it'll still have vocabulary but
04:04 : usually the vocabulary is very generic
04:06 : such as just very generic sounding
04:11 : variable names like X or I don't know
04:15 : num for number or string billions
04:21 : objects oh it's all like code or speak
04:25 : other vocabularies have to do with the
04:28 : domain of code and yeah coders can talk
04:32 : in that language but not the business
04:34 : domain people and business domain people
04:38 : could care less about reading code
04:40 : that's ignorant because it doesn't they
04:42 : can't relate to it
04:44 : so that would be ignorant code versus
04:47 : savvy code savvy code very
04:50 : business-oriented
04:51 : of business knowledge economical very
04:55 : little or none in terms of business
04:58 : knowledge clever what do you mean by
05:00 : clever clever code is very intricate
05:04 : there's a lot of like if you are to you
05:10 : had a piece of code I don't know that's
05:13 : perhaps very nested like it's very dense
05:18 : in terms of content and if you drew a
05:22 : diagram it is so complex first of all
05:24 : then you need to draw a diagram in order
05:26 : to hope to understand what's going on
05:28 : and if you did draw a diagram that
05:31 : diagram might be very complex and that's
05:34 : what I mean by clever so clever piece of
05:38 : code is kind of complex and take someone
05:42 : out of mental power just to understand
05:47 : what's happening in that piece of code
05:50 : and the opposite of clever and thus
05:54 : which is why I talked about like the
05:58 : cleverness the ingenious like it seems
06:00 : like like someone has to be a nuclear
06:03 : physicist to understand this piece of
06:05 : code like the person who invented the
06:07 : world
06:07 : why well for example was a nuclear
06:10 : physicist worked at CERN that's what I
06:14 : heard I hope that might be true about
06:19 : web programming today very smart person
06:24 : to understand it but hopefully you can
06:26 : do better than that
06:27 : and not everything needs to be super
06:30 : complicated but the opposite of clever
06:32 : with I'm gonna call that dumb dumb dumb
06:38 : code is code that's very straightforward
06:40 : so dumb code looks like
06:49 : it's just like one after another and
06:51 : then as you are reading you can sort of
06:54 : it's like a novel that you can sort of
06:56 : go from the beginning to the end in one
06:58 : seating and you felt like yeah okay I
07:02 : got it it's not anything that requires
07:05 : mental reflection like days like you
07:08 : read some books under that you it
07:09 : requires days of mental reflection
07:11 : afterwards that's a very dense book dumb
07:14 : is not that okay
07:16 : so that that's what I mean by clever
07:19 : versus savvy and with any piece of code
07:22 : you can categorize it in both of these
07:25 : ways so so it's possible to categorize
07:30 : any piece of code artifact in and put
07:34 : them in one of these four quadrants
07:37 : now it's relative it's not it's not like
07:40 : there's a gradient as well like if you
07:44 : can think of this as the plane where you
07:46 : plot points as well if you want it to be
07:48 : more detailed about like it it could be
07:50 : here versus here versus here etc so but
07:55 : roughly speaking we can clicks code in
07:58 : one of these four quadrants so then like
08:01 : what is what does this mean like what
08:03 : kind of code lives in this first
08:05 : quadrant it's both dumb and ignorant huh
08:11 : pure functions okay
08:16 : are you saying fewer functions cannot be
08:19 : clever no I think that's too broad a
08:26 : statement
08:27 : let's OC is object in low - is our is
08:35 : array although you haven't seen its code
08:38 : so we'll dig up this piece of code I've
08:42 : dug it up before it's not that
08:44 : complicated so this is a piece of code
08:48 : that is both in my opinion both very
08:52 : dumb and very ignorant it's it's it's
08:57 : very ignorant because there's no
08:59 : business domain language in here it's
09:01 : all code speak it's all talking about
09:03 : objects and types and functions and
09:06 : values business people can't read this
09:10 : they don't want to so it's very ignorant
09:13 : it's very dumb because the logic it
09:16 : embedded here is very straight for and a
09:18 : very little content in the first place
09:21 : there's it's two lines function can
09:23 : anyone else think of other examples
09:25 : perhaps from Oracle base that don't you
09:28 : would put in this first quadrant here
09:32 : get like low - get is actually somewhat
09:36 : clever it's a recursive function I've
09:41 : actually put it over here low - low low
09:48 : - FP get and a set set is even more
09:54 : clever that that I love this function
09:57 : let's go to this quadrant there so
10:00 : examples of code artifacts that are both
10:03 : clever and ignorant I used two examples
10:08 : what are some others sequel yeah yeah
10:11 : yeah sequel database psycho engine of
10:16 : some sort or any kind of query language
10:19 : equal for those who don't know equal is
10:22 : are in
10:23 : query language for curbing data but but
10:27 : really I would actually put most of
10:29 : lodash in this category except for
10:32 : probably these fuels that are like this
10:35 : in this quadrant there there's a lot of
10:36 : low - that go in here
10:38 : rxjs definitely belongs in here if you
10:45 : are a framework you probably should
10:47 : belong here
10:48 : so I would put oh I'm not saying the
11:01 : code that uses this I'm saying the code
11:04 : within the code within set if you crack
11:08 : open low - set and you look at it it's
11:11 : actually very interesting no because you
11:21 : could have you have very savvy code that
11:24 : just happens to use the low - functions
11:28 : but it you could have a lot of business
11:31 : terms in that code store is something
11:34 : that we engineered here if it's a
11:36 : concept that we prototype and we're
11:40 : using now in our apps for those who
11:43 : don't know and VMs is the view model
11:47 : store which is sort of an extension of
11:49 : the store
11:51 : Oh soso VMs VMs you're saying yeah
11:59 : so the store to rate rose point I think
12:03 : store is actually lower on the
12:06 : cleverness scale compared to something
12:10 : like rxjs or react or angular or a low -
12:16 : and that's actually why I like it so
12:19 : much even though it's clever it's
12:21 : actually towards towards this way in the
12:25 : clever versus dumb scale yes top is
12:39 : ignorant so everything we've done so far
12:41 : is alright except for the VMS the old
12:45 : model store so what is something that
12:48 : goes into this bottom left quadrant some
12:51 : examples a screen in our app that's very
12:54 : straightforward do we have any of those
12:59 : the page config screens like the JSON
13:02 : configuration if we call that code
13:08 : page can fix the JSON code not the
13:17 : engine that makes the JSON code work
13:20 : just a JSON code that that is an example
13:23 : of dumb and
13:31 : clever and savvy you keep jumping ahead
13:34 : on from that from the perspective of the
13:44 : JSON programmer hopefully he doesn't
13:46 : have to do a lot of for people who don't
13:53 : know what this is we built an in-house
13:56 : JSON based language for coding pages and
14:02 : and form these pages and tables so you
14:05 : can code an entire page using only JSON
14:08 : and that's what what we're talking about
14:11 : and we build the engine that supports
14:13 : that JSON based language for building
14:15 : pages so I'm saying when you're coding
14:19 : up a page like that in just JSON you're
14:23 : not having to do a lot of clever Ness
14:25 : it's mostly business verbage that you're
14:30 : using in the debt JSON language because
14:32 : you're you're talking about particular
14:34 : fields you have and the relationships
14:36 : between the fields and those field names
14:38 : are in in business verbage yeah any
14:42 : other example do we have any other like
14:45 : dumb screens like code that's very savvy
14:49 : but very dumb like it's very
14:51 : straightforward to read widget
15:12 : okay well that was hard be submissive
15:20 : generic table okay generic we have a
15:25 : reusable component called generic table
15:29 : for rendering rendering a table in the
15:33 : UI and also allowing editing of the data
15:38 : and but that's we're really we really
15:42 : intend for it to be here but at the
15:45 : moment it's sort of straddling the
15:48 : middle what work plan is probably for
15:55 : the Churchill form which is a generic
15:58 : component it has a lot of cleverness but
16:02 : also has a lot of business logic in
16:06 : there
16:06 : that's the savviness we have a work plan
16:11 : okay given this sort of like this this
16:15 : worldview that I'm overlaying on top of
16:19 : sort of code that how we look at code
16:23 : what's what's the insight I'm trying to
16:26 : give I'm basically saying the code
16:31 : that's in this quadrant the first
16:33 : quarter dumb and ignorant that's a math
16:36 : as in it's probably not ideal but it's
16:41 : also not hurting us like that the
16:43 : presence of code in this quadrant
16:47 : probably signals that there's something
16:50 : about the underlying platform we're
16:52 : using that is lacking and therefore we
16:56 : need to sort of fill in some gaps
16:59 : the platform being say JavaScript or
17:03 : angular or types there's no quick fix to
17:08 : that problem we're just gonna live with
17:10 : it and it's not hurting us too much yeah
17:16 : a lot of the lodash stuff that's not
17:20 : very clever like I just give two
17:24 : examples here did might be more just
17:27 : like something I could have one line
17:28 : this but I chose to use lodash because
17:31 : it's more meaningful when it says is
17:33 : object then type of object or this and
17:38 : that and the other thing is so it's
17:40 : helping you but not that much the ROI is
17:43 : not quite as much as if you were to do
17:46 : something more than that
17:48 : and then I'm saying code that's in the
17:52 : second quadrant here clever and ignorant
17:55 : that's good we want that and code that's
17:59 : in the third quadrant dumb man savvy
18:03 : that is also good we want that and I'll
18:08 : come back and explain why and but the
18:11 : code that's in the fourth quadrant we
18:14 : should try to avoid this so the this is
18:17 : really my main point of drawing all
18:21 : these things out I'm gonna try to
18:24 : perhaps rationalize why you want to
18:29 : think about the sources of change in a
18:34 : software application where when change
18:37 : is happened where is that coming from
18:40 : well usually that's coming from the
18:43 : clients of the software they require the
18:48 : software to have this feature and have
18:51 : that feature that the clients are
18:54 : driving the change and because your
18:57 : clients are not software people the
19:02 : verbage they're going to ask the change
19:04 : in is going to be in business
19:07 : language so they're they're gonna
19:09 : they're gonna ask for things to change
19:13 : in terms of things that they understand
19:16 : which is based on the business language
19:18 : so if all of your code that has business
19:24 : verbage in it is in the third quadrant
19:28 : that's gonna make our lives easier
19:32 : because of that code because code in
19:35 : this quadrant is easy to understand
19:38 : because it's dump you can read through
19:41 : the code and not be confused so it's
19:46 : like oh yeah I know what that's doing
19:48 : and then to satisfy my clients request I
19:51 : just have to change this to be this and
19:54 : I'm done if you have a lot of if you
19:58 : have code in the fourth quadrant that's
20:01 : affected that's a problem because that's
20:04 : very clever so it requires a lot of
20:07 : mental power to sort out what's going on
20:10 : in the first place to be the point of
20:12 : being able to modify it confidently and
20:18 : successfully to make it do the new thing
20:21 : that the client wants however if the
20:24 : clever stuff were ignorant if they were
20:26 : in the second quadrant then they're sort
20:29 : of insulated away from what the clients
20:33 : are directly asking for in a sense
20:36 : because it's so ignorant whatever they
20:38 : say they are going to directly affect
20:41 : the savvis the stuff in the bottom
20:44 : section but not directly in the ignorant
20:47 : section so there's a sort of insulation
20:51 : due to the ignorance of that code but
20:54 : the fourth quadrant is the quadrant of
20:56 : pain because they ask for the things to
21:00 : change you have to deliver it tomorrow
21:02 : and you're struggling to understand the
21:06 : code that's there in the first place so
21:08 : that's the rationale
21:09 : geminus yeah any question and any
21:14 : comments from anyone at this point
21:21 : that's a good point
21:23 : your your client might actually be
21:25 : technical people they might want in on
21:29 : the design process or I guess that's
21:32 : sort of like your client is like a
21:35 : developer actually which does happen in
21:38 : real life like we might be a company
21:41 : that makes developer tools for example
21:43 : and then our client is actually
21:45 : developers their business domain is code
21:48 : so they're talking about stuff like
21:50 : strings and arrays and models and stuff
21:54 : like that that that's an interesting
21:55 : scenario as well but we don't
21:58 : necessarily have that well we do have to
22:01 : work with some people that is deep in
22:05 : the Dakotas what are we really trying to
22:06 : do as a software team if our goal is to
22:13 : make software that's more maintainable
22:16 : and also work better as a team in
22:20 : general because your team is gonna
22:22 : screech to a halt if you don't have
22:25 : maintainable software so basically you
22:28 : want to identify the code that's in here
22:31 : and then you want to try to break it
22:35 : apart and move one chunk of it into here
22:38 : and another chunk of it into here
22:42 : yes so so whenever you're trying to work
22:45 : on improving the software that's already
22:48 : there you might call that legacy code or
22:53 : technical debt or however you want to
22:54 : frame that if you want to improve the
22:57 : quality of the software this is what
23:00 : you're trying to do come up with a
23:02 : clever way to break it up into two
23:06 : slices and you move one of the slices to
23:09 : the second quadrant
23:12 : the other slice into the third quadrant
23:15 : okay thank you

Clever vs Savvy - Part 1

Transcript