Hacker News Reader

Comment by cl8ton | Hacker News Reader

I am one of the older'ish programmers

One day a young programmer flies into the office talking about the new MVC frameworks and how it would revolutionize how we write programs.

He was mystified when I explained how MVC works to him without even looking at what framework he was ranting about.

He asked how I knew? I told him it was invented at PARC in the 70's and I've used them starting in 1995 with the MFC C++ for document view. This was more of a reintroduction of an older methodology suited to web.

Older developers also get to see the linear progression of current technologies and how they arrived at current day implementation. What worked and what didn't work.

View on HN

There is very little new under the sun, at least if you're talking about what's going on in the mainstream rather than what's happening in the theoretical space (with GADTs, dependent types, etc). It's been a very long time since I saw software development technology that wasn't just the application of something developed in the 1960's through 1980's. "Oh, Java has lambdas now? Here are the design patterns that take advantage of them. They were well-understood in 1975." "Oh, Ruby has generational garbage collection now? Here is the paper describing the performance characteristics of that--from 1984."

I think TypeScript is actually one of the more genuinely interesting developments recently--it's state of the art circa 1996 (Abadi and Cardelli's "A Theory of Objects.")

This is not really a dig at software development. There is, for example, very little in aerospace that has changed more than incrementally since the 1970's. But people have this perception of software development as this rapidly changing field where knowledge becomes obsolete in the course of a couple of years, when in reality the basic principles are really quite well-developed and mature. Keeping up with software development technology is mostly a matter of figuring out how people have gussied up old ideas with new syntax.

All of the interesting stuff now is domain specific. Algorithms for self-driving cars? Yeah, that's something that we didn't have 20 years ago.

And GATDs and dependent types are pretty old too by now. E.g. Martin-Löf Type Theory was first published 1971-1972. :)

That's interesting. What is it about Typescript specifically that you think aligns it with "A Theory of Objects"? (not expecting a thesis in your answer)

Abadi and Cardelli lay out an object oriented calculus with object literals and structural object types. Typescript's object literals and structural object types seem very similar.

Thanks for that.

Yes. Twenty and even thirty years ago we did have essentially the same ML algorithms used in self-driving cars today, just not enough computing power to use them effectively (I think SVMs are only about 20 years old but neural nets have been around in some form since the late '50s) There is much to learn from those who have gone before.

Yup, it has happen to me too. I have around 20 years in software development.

Around 9 years ago...

Me: We can abstract most of that data access with a simple table oriented orm. Not all of it, but a good 80%.

Younger Developer: Nope, That is what record sets are for, it is right here in the Microsoft docs (pointing to the ado api)

...

5 years later, he had moved to a different company, and I received a phone call from him: Just called to say that now I understand what you wanted and how it is simpler.

Me: Then I failed to communicate properly. What could I have say that would have make you give it a chance?

Him: Nothing, I needed to learn it the hard way.

Experience teach.

A few months ago, working with very young dev. team.

Tech lead (late 20s): We need to do all using TDD, if it is not TDD we don't touch it.

Me: Agree, but TDD is one of several tools to DDD; the developer needs to understand what its building, not just that all checks are in green.

Tech Lead: The devs only need the proper test.

Now they drop the core calculations from the project because they couldn't understand them by just looking at the tests, I got task with that, and a salary increase. And some how the tech lead think it is unfair because he "gets" Git and Resharper much better than me.

Aaawww, the luxury of being young.

Respectfully, I have to disagree with your assessment that age is the contributing factor here. I'd argue that pragmatism is what you're seeing.

Pragmatism might correlate with experience, and experience is only gained with time (thus age), but I'd be careful about jumping to that conclusion. There's nothing beyond assumption (this article included) that says younger people are always, or even usually, less pragmatic.

[Edit: That said your tech lead is an idiot if he thinks he should be paid more for a better understanding of technologies tangential to his core goal while he has a weaker understanding of the goal itself.]

Respectfully I submit that we might not be in a big disagreement here.

Pragmatism is a very high contributing facts. My personal experience has not show me IF pragmatism is better. Instead experience has show me WHAT is worth to be pragmatic about.

My observation is that experience leads to emotional maturity, which leads to better pragmatic choices. It is the fear to make mistakes, hence the industry coined "Fail fast". It is the ego boost of "I just learned this cool technology", as opposed to "If this is so cool, hasn't been invented before? How did it look like?" (Another post showed an example with MVC). It is the "I own the world" feeling that we all have as young adults. And there are some "bad" apples in the VC culture that can exploit that.

[And I don't think the tech lead is an idiot, he is where I was 18 years ago]

> Respectfully I submit that we might not be in a big disagreement here.

I like it when that happens. :-)

> And I don't think the tech lead is an idiot, he is where I was 18 years ago

I shouldn't have phrased it that way. The idea that you're more valuable than someone who can solve a core business problem when you can't is idiotic. He himself is probably not an idiot - we've all let our emotions get the better of us.

Your first younger developer sounds awesome. I've worked with people (some older, some younger) that refused to learn from their mistakes.

All developers make mistakes -- "I built this because I was worried about X but I should have been worried about Y." The problem is making the same mistakes over and over.

One of the few joys of life is discovering new things, and to lead a fulfilling life one must discover new wrong ways of doing things. :)

There is indeed very little new under the sun. A lot of stuff are just repackaging of old ideas. I remembered reading the Self language paper and thought it was cool as heck, objects all the way, no classes. Then later Javascript came along with its prototype and people thought it's the shit. I was like, whatever, it has been done before.

If I remember correctly Brendan Eich was originally promised that he could build Self for Netscape, but later due to various reasons had to change the syntax and all that.

Also the popularity of javascript in the geekdom is in my recollection fairly recent phenomenon. It became interesting only after ajax, before that it was a leper like php.

Wow, when kids don't even know that the latest fad was originally popularised by Apple, we have entered postmodern hipsterdom.

Are you referring to MVC in iOS programming? I would attribute the current popularity of MVC to Rails.

There were lots of MVC frameworks for web development that predated Rails. For example Struts.

And MVC was talked about a lot as a design pattern for Web apps before Struts was started - I remember we rolled our own rather nifty one in 2001 or so and failed to get our execs to open source it so it died when the team withered away after our acquisition.

Struts was released in 2000, so it predates yours. However I know people who were using MVC as a design pattern in Perl in the 90s.

MVC comes from the Smalltalk world. http://c2.com/cgi/wiki?ModelViewControllerHistory

But it is quite different from web MVC that was popular with web frameworks like Struts.

I haven't looked at the more recent javascript frameworks, but my intuition tells me that they are probably going back to the original MVC pattern. (Observer pattern to notify views when model changes)

Said history was referred to in the top level comment that started this discussion.

I've had as a low priority idea for a few years implementing a web MVC framework that is somewhat closer to the original inspiration - not because it makes sense for generating HTML (though it actually works quite well for that) - but because it would reduce the impedance mismatch with a client-side library using the original MVC pattern.

I have not actually done this for a number of reasons. The top one being that when you scratch an itch that you think someone else should have, it is much less fun and likely to work right than when you scratch your own itch. But it still bugs me that nobody has written one that works like I think it should...

What I meant was that Struts wasn't the first mention of MVC in relation to web apps - lots of people were rolling their own frameworks in 2000/2001 and MVC was a common buzzword.

As most developers in those days hadn't started with web apps it wasn't surprising that the term MVC was used - e.g. we were writing MVC C++ wrappers for simulations in 93/94 and Java MVC simulations in applets from '95.

I agree that Struts wasn't first on the web.

But a 2000-2001 handrolled MVC framework does not qualify as "pre-Struts".

LOL ... I am old! I learned MVC developing MFC apps back in the 90s!!

I resemble that remark... I'm also old.

I was writing RDBMS in dBase, compiled with Clipper, on MS-DOS, in the early 1990's, when I was about 30 years of age. (I was the "old guy" 20 years ago)

Wow! The decades just keep passing by with increasing relevant speed.

Clipper! yes that was the first language I used professionally.

I tell the younger guys that before IDE's and flat memory models were around, when you programmed, you went looking to buy three things.

1. Language/Compiler 2. Linker 3. Editor

And you will spend just as much time in the linker file (segmented memory back then) as you did coding.

The newer tools are just light years ahead of where they started back then. Its one area I will never look back at and wish it was still around.

It is a well-documented fact that time passes as a percentage of life lived. There is far less relative distance between decades for me now than there was between my fourth and fifth birthdays. I have kids who have kids, and they weren't around for that last episode of MASH (left out the asterisks for formatting reasons) that I'm pretty sure (in what's left of my mind) was first aired a couple of years ago.

A friend of mine told me this longer ago than I care to be reminded of. "The older you are, the faster time goes. The telescoping effect can be quite startling."

Reply on news.ycombinator.com