intertent.net 3.0

I just saw Jeff Atwood’s article on the subject of programmer experience.

Programmer Experience

Jeff questions the correlation between a programmer’s experience and their programming ability, he also talks about company’s who won’t hire anyone without experience in the exact system they need the developer to work on.

Software development is a funny thing - while there is definitely technical knowledge involved, there is also a fair bit of “secret sauce” required to get any non-triivial project off the ground.

When you get past a certain (very small) size, the structure and design decisions you’ve made when first writing your code make a big difference. So although a novice programmer may be able to write a function that opens a window. They will most likely drown if they try to write even a simple text editor.

The best example of this is with 3d games - getting a spinning cube to display is relatively simple. But writing a 3d game requires quite a bit of design skill.

It’s in this facet of programming that experience and “gut” feelings become very useful. A programmer without a certain level of experience will most probably not be able to design a large system on his/her first try.

Another aspect of this is that of good judgement - there are some people who just have a knack of consistently making the right decisions. In my book this attribute is much more important than years of experience or any other measurable factor. What troubles me, is that I don’t know if you can really tell something like this from a resume.

Relevant Experience

I’m of the opinion that a reasonably experienced programmer - one whose smart, knows a couple of programming languages, and has done a couple of reasonably sized projects. Is capable of becoming productive in most environments fairly quickly.

Therefore, I think any company limiting itself to people with “5 years experience in XML” are needlessly narrowing their pool of candidates.

By and large, the only real exception I can see to this is when a company needs someone to develop a system now [as in yesterday], and so can’t afford the 3 months or so it would take for a new developer to learn the environment. It’s important to note however, that even a programmer with loads of experience in the specific field is likely to take a while to come to terms with an existing codebase, so the perceived win may be smaller than what one would think.

Dion Almaer posted a great interview with Steve Yegge concerning his port of Ruby On Rails to Javascript - “Rhino On Rails“.

Rhino On Rails sounds incredibly cool, but I wanted to talk about some of the other stuff Steve mentions in the interview.

Some people think that the web needs a new architecture:

Web development is unimaginably obtuse nowadays.

• HTML wasn’t designed to do the pixel perfect presentation that everyone seems to want out of it, so you have to hack your way around.
• Web browsers aren’t compatible with one another, and as a browser vendor you have to emulate bugs in the market leader in order to render sites correctly.
• Progress is slow - 5 years in between IE 6 & IE 7, standards bodies that sit on non-relevant standards for years (e.g. The CSS Working Group).
• A large number of uninformed developers, which is magnified by the leeway you have writing incorrect HTML & JS while still having it render with no warning in the browser.

It’s tempting to think that maybe some uber-hacker can come along and design a “new standard” that everyone will suddenly start using. Something clean and fun to use, which solves the real problems people have.

Unfortunately, I don’t see this happening any time now. For something to replace the HTML,CSS,Javascript trifecta it’s going to have to overcome one heck of alot of legacy. The web, as currently defined is ubiquitous - it’s in our PCs, phones and toasters. For something new to take hold, it would have to be at least 10x better, and have very good PR. It would also have to be an OPEN standard, as even a very large company wouldn’t have the resources to write enough good implementations.

Unfortunately I don’t see any of this coming.

Server Side Javascript:

Steve says that after you get used to it, writing server side javascript isn’t that bad.

What bugs me about client side Javascript is things like:

a = function(){return 1;}
()

Because Javascript tries to be smart and inserts semicolons for you, in the above example a is equal to 1, and not an anonymous function that returns 1 as you might expect. Oh wait… That’s actually 1.0000f - Javascript currently only supports floating point numbers and has no int type. It makes me cringe thinking of the subtle bugs you can encounter due to quirks like this.

I hope they’ve solved that in Rhino and/or EC4.

Steve is excited about D:

That really brought a smile to my face. D looks like it’s shaping up to be a worthy successor to C++. It still has a long way to go, and allot more support to garner, but if Steve’s excited about, there appears to be hope .