|
I would suggest that you inform your maintenance team about the subtle but crucial difference between "simple" and "simplistic" 
|
|
|
|
|
There are no 20 lines of code that are simpler than 1 line of code (assume that it fits in 80 characters margin), unless 15 of those lines are comment. The idea of dynamic languages is the model, not the structure. If you want a structure, you can stick with java - there zero lines of actual code require at least 20 lines. In dynamic languages the complexity of the task/model (which should be understood if you try to understand the code) is directly related to how much lines you use - less lines = much simpler (but that does not apply to unformatted code).
|
|
|
|
|
When I started my first programming job the team were using legacy C libraries from another company that were old, clunky, terrible to use, full of bugs and required a lot of typing. When creating new programs they copied forward an existing one that did most of what they wanted to do and added a load more code, commenting out any bits they didn't need just in case when it copied forward again the original code would be in there. It was a bit of a mess!
I learned their libraries inside out, but instead of saying, all your code is awful why don't you do X, I set about rewriting them a tiny bit at a time and slipping the code into the programs; at first people didn't really get it but after a while (and it was a while) they realised ways of doing things had the same/better resulst and took less typing and started doing it in the different way.
Some people don't like change and will cling on to old methodologies and code because it's either what they know and can do, they don't want to put the effort in to learn anything new, that or they think you are usurping them in some way and if they "allow" you to "update" things it implies they were crap in the first place for not doing that ages ago. I'd recommend starting very small or moving to a department/company that likes moving things forward at a faster rate.
tldr;
I'll let future me worry about this.
|
|
|
|
|
It sounds like what your company really needs are regular technical sessions whereby any developer can do a quick demo of a particular technology or framework that they think is useful.
Getting all the developers together with a 10-15 demo showing what Knockout can do, how it reduces the amount of code you have to write, how easy it is to make things work, and how it would make future maintenance much easier would get any professional developer on board.
If it doesn't, I'd start looking for another job.
modified 31-Aug-21 21:01pm.
|
|
|
|
|
This sounds like a good idea!
Although this won't really work in our case. I work at a consulting company (the best one I have been at so far!  ) so we are all dispatched to different client and do not really work together!
Although some developer did mention they would like some technical meeting like that to happen... I wonder...
|
|
|
|
|
One line of framework code is not necessarily simpler than 20 lines of lower level code if people don't understand the framework, if the framework is complex or a maintenance headache in itself, or if the framework doesn't naturally fit the problem and you end up with 30 extra lines elsewhere to work around it or set it up. Simplicity is not just a measure of code size, it's a measure of understandability, and if a seemingly simple framework call is doing a lot of stuff that people don't understand, it isn't simple.
(To use a different example, .Net Task spawning code looks really simple. new Task(lambda).Start(), what could be hard about that? But to understand that one line, you need to understand threading, thread pools, synchronisation/locking concerns, marshalling UI calls to the UI thread and so on, and if you don't, it is harder to maintain Tasks code than it would be to maintain the longer but more explicit manual thread management code.)
Those things may not be true in your situation with Knockout.js, although the first one clearly is, but it is the responsibility of the person who wants to bring a new framework to a project to introduce it to the team and make sure that it really does make the team's life better, not just yours.
You mention lambdas and generics in your edit. I've been 'that guy', particularly with generics – but I won the argument because I could explain why the generic code was simpler and that it was worth the investment to learn.
|
|
|
|
|
You make a good argument ...
But it doesn't quite apply in my case.
I don't even work with this maintenance team. Have no regular work-tech meeting with them. They just had a look once at the repository.
When I meet them informally, say for lunch, I hear a lot of complaints... I even say let's "sit together" but guess what? "there is no time"!
As for the people I work for? They are all converted by now!
|
|
|
|
|
Well in that case if your PM is onside, just tell them they're wrong, that they need to go learn about the technologies used in the application they're maintaining, and direct them to the PM if they start an argument about it.
|
|
|
|
|
Part of working successfully as part of a team is consideration for others. If you're not able to calmly and logically persuade them to your point of you then I am afraid you will have to swim with the tide. Regardless of who is 'right' or 'wrong' or how much it costs for either party.
|
|
|
|
|
Well this is difficult in this case.
I am not hit by my team mate. But by some remote team who had a look at the source code and with which I can't talk!
I already asked them: "let's sit together and 'fix' the code" but... "there is no time"
|
|
|
|
|
It can take awhile to change people's long held views. But good luck 
|
|
|
|
|
OK, how does your 1 statement knock out codeline look like?
Show, don't tell !
|
|
|
|
|
Maybe 20 lines of jQuery was bit exagerated.. 
But the latest cookie knockout extension I wrote is quite concise, expressive and helpful.
short version
<div data-bind="cookie: { name: 'visibleColumns', data: visibleColumns() }">...</div>
long version (where I inline visibleColumns())
<div data-bind="cookie: { name: 'visibleColumns', data: tableOption.aoColumns.enumerable().where(function(x) { return x.koVisible; }).select(function(x) { return x.koVisible; }).array() }">...</div>
I have a datatable extension, which take some standart datatable option, and some extra option like an (optional) 'koVisible' property of type ko.observable<boolean>() (TypeScript syntax), to automatically hide / show columns.
This statement collect all the koVisible flags (on columns which have it), initialize them from a cookie if it exists and automatically update the cookie if any of them changes.
modified 2-Jun-14 9:17am.
|
|
|
|
|
How about:
Keeping It So Simple Makes Your Application Run with Superior Efficiency
Or, for the 'merkins:
Keeping It So Simple Makes Your Application Stable and Solid
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
Think I can use the acronyms!?!
|
|
|
|
|
Keep fighting the good fight. If you are using knockout I assume you are also doing MVVM. MVVM is all about maintainability and simplicity, not to mention KISS/SOLID.
A couple points for you:
A) Just because you don't understand some code doesn't make it bad.
B) Those opposing MVVM and enabling frameworks in todays development world are analogous to people in the year 2000 something still calling OO a fad.
Regardless implementing a thoroughly tested and production ready framework over writing a bunch of roll your own code is usually going to yield less bugs and more bang for the clients buck.
|
|
|
|
|
Rowdy Raider wrote: Regardless implementing a thoroughly tested and production ready framework over writing a bunch of roll your own code is usually going to yield less bugs and more bang for the clients buck.
That is a good repartee! :P
|
|
|
|
|
Rowdy Raider wrote: Keep fighting the good fight. If you are using knockout I assume you are also doing MVVM. MVVM is all about maintainability and simplicity, not to mention KISS/SOLID.
Indeed!!!
|
|
|
|
|
Simplicity is always in the eyes of the observer.
The question is which eyes to care ?
Maybe you can improve the eyes of the maintenance team by training them. But it can only work if they agree to. (Lot's of developers hold grudge against MS will despise knockout and typescript 'just because')
On the other hand you can just ignore them. You are the one getting things done, and stackholder pushing your project don't care about whether or not maintenance team can take it.
The root of the problem is that HR department indirectly pushes "KISS for everybody" so they can hire more easily. (Which is reflected by the level of your maintenance team, that in your case is clearly incompetent for .NET development)
But, at the end of the day, you are the one making another's life better, and in the position to dictate what to use.
Final answer : train them, and if they refuse to be trained, ignore them and get stuff done.
|
|
|
|
|
Super Lloyd wrote: Hey it's true, they look at it dumbfounded and say "it's unmaintainable" so the average programmer can't fix it, indeed!
If the average programmer can't maintain it then it isn't maintainable. The only subjective part of that is whether the average programmer can maintain it.
Super Lloyd wrote: What I sell to the customer, is not just humanely doable in a reasonable time frame with just jQuery.
Why are you selling to a customer?
And why is maintenance not part of the sale?
Super Lloyd wrote: Very tiresome!
Myself I consider it a compliment when someone says that my code is easy to maintain. That is because I realize (as been proven by numerous studies) that the initial cost to develop something can be anywhere from 50% to 1% of the overall cost to maintain it. And being a professional working in a business (not a hobbyist) and one who likes to be paid I keep the primary goal of a business, to make money, always foremost in my mind. Thus I at least make the attempt to help keep maintenance costs down.
|
|
|
|
|
jschell wrote: Why are you selling to a customer?
And why is maintenance not part of the sale?
Let's say I sold Facebook. It's not Facebook of course, but it is a little bit more dynamic. Things are popping here and there, there are multiple dynamic view of the same data. Mmmm... in an other word it's a SPA (single page application)
Now I just can't develop a SPA with JUST jQuery. At the very least I would need to write a library of utilities. They are against (home made) libraries. If my code is not all on the page is too complicated. And this... is... just... beyond... my capabilities...
|
|
|
|
|
For me Keeping It Simple means that anyone in the team is able to understand, maintain and modify what i'm doing without too much trouble (or too much asking), and/or that it have less probability to go wrong, from this point of view, using one line of Knockout would be simpler, because it could fail less often than 20 lines of hand rolled jQuery that do DOM manipulation, and the maintenance team just need to learn a new tool to be able to maintain it!
|
|
|
|
|
There is also the fact I just CAN'T develop what the customer now want (my prototype, a nice SPA (Single Page App) with JUST jQuery and NO home made library.
|
|
|
|
|
The problem isn't KISS. The problem is people who are not keeping their skills up to date. If they were always keeping up to date, they would take some time to evaluate the issue and see that things have improved and they can avoid all those lines.
It is a problem to maintain because it looks like magic to them and they don't know the incantation. Chances are, they program by finding some code that does similar and modifying it as best they can. And there is a better than even chance these people don't really get what is happening there either.
|
|
|
|
|
You can routinely replace 20 lines of jQuery with one line KonockoutJS?
Interesting.
modified 20-Oct-19 21:02pm.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
