DevTalk.net - Latest Comments

Re: Chained null checks and the Maybe monad

Giacomo S. S. — Sun, 24 Mar 2013 16:52:26 -0000

Great post, also if I prefer I simpler Maybe monad (like this http://blog.ploeh.dk/2011/02/0....

In 2013 after the diffusion of many FP langs and after a lot of langs like C# got great functional constructs... there's still people ask "why?!".

Why we should avoid boilerplate unreadable code that check nulls? This post and also the one quoted by me, explain "why".

I've wrote too a lot of code that check a lot of times for null, I admit. But once understood I've shame of this code and for the future, you'll never see a null check in my code (except for parameter guard clauses).

This is why.

Re: Chained null checks and the Maybe monad

Tormoz — Fri, 14 Dec 2012 17:54:40 -0000

I had a colleague who wrote his own Maybe<t> monad, and my initial reaction was, Why? And: I can't read this. It took me a while truly to appreciate the flexibility and improved readability it provided. Yes, the terseness can be an initial barrier to understanding, but one I "got it" I really appreciated it. Sometimes it's a matter of which verbs we select: some people like "With," for example, and others won't.

In any event, thanks for making a complex topic very understandable -- and for providing repeatable examples too.</t>

Re: Chained null checks and the Maybe monad

Dmitri — Sat, 02 Jun 2012 03:51:32 -0000

I guess if all you're doing is checking for nulls, it is maybe a bit more meaningful. However, that's not always the use case. For example, you might have a Nullable<T> in there, or maybe a struct. In this case, calling the function When() makes a lot more sense.

Re: Chained null checks and the Maybe monad

MichaelFreidgeim — Fri, 01 Jun 2012 19:39:26 -0000

The extension name With is not meaningful for me, I like IfNotNull from http://smellegantcode.wordpres...

Re: ReSharper SDK Adventures Part 2 – Math.Pow Improvements

Dmitri — Thu, 10 May 2012 17:53:02 -0000

Yep, I imagine this would work too, though it's a slightly less comprehensible alternative. There's a hundred ways to skin the cat, as they say :)

Re: ReSharper SDK Adventures Part 2 – Math.Pow Improvements

derigel — Thu, 10 May 2012 17:09:26 -0000

bool isOnMathPow = false; var r = element.InvocationExpressionReference.Resolve(); var m = r.DeclaredElement as IMethod; if (m != null) { isOnMathPow = m.XMLDocId == "M:System.Math.Pow(System.Double,System.Double)"; }

Re: ReSharper SDK Adventures Part 1 – Math.Pow Inlining

Alexander Buryak — Fri, 04 May 2012 13:28:52 -0000

Very nice overview, thank you!

Re: Chained null checks and the Maybe monad

Rodrigo — Thu, 03 May 2012 16:26:55 -0000

You should maybe re-read this.

Re: Realtime Programming

Dmitri — Wed, 11 Apr 2012 06:57:10 -0000

Nice!

Re: Realtime Programming

New User_ — Mon, 09 Apr 2012 04:05:10 -0000

getting more out of FSI REPL see http://code.google.com/p/fseye...

Re: Local Inversion of Control

Dmitri — Fri, 06 Apr 2012 13:15:26 -0000

Thanks, glad you like it.

Re: Chained null checks and the Maybe monad

Dmitri — Wed, 21 Mar 2012 09:39:00 -0000

I think you're missing the point. A single null check is no problem, but lots and lots of nested null checks take readability to zero.

Re: Chained null checks and the Maybe monad

Computersade — Wed, 21 Mar 2012 08:52:31 -0000

Can someone explain to me how this:
if (person != null)
is " fairly unreadable (and un-maintainable) code"? It reads as "If the person object is not null". Seems perfectly readable (and maintainable) to me.
Even in long hand "if open bracket the person object is not null close bracket" it still makes sense.

Now this:
this.With(x => person)
is what? Readable? Maintainable? It reads... umm, not sure, erm let's look at the "with" method... ah, when evaluating, using the with method, the object "this" - is the result, returned into variable x, equal to or greater than the person object?!? Is that right, erm, maybe I should have another crack at it...
In English it actually reads "this dot with open bracket x is equal to or greater than the person object close bracket". What does that mean??

(Nice bit of code though)

Re: Introducing MathSharp – Convert Math to Code

Art Scott — Sat, 18 Feb 2012 13:36:01 -0000

Way!!

Re: Local Inversion of Control

rob sutherland — Fri, 17 Feb 2012 11:49:26 -0000

Really love the obj.IsOneOf("a", "b", "c") extension. Big fan of extension methods in general and tend to create a lot of them, but haven't done this one yet.

Thanks for the post.

Re: Local Inversion of Control

Dmitri — Fri, 03 Feb 2012 02:51:03 -0000

I don't have a public package with my methods. Hmm, perhaps I should publish one...

Re: Local Inversion of Control

Igor Sidorov — Fri, 03 Feb 2012 01:28:43 -0000

Can I get package with your extensions methods somewhere?

Re: Web Mining via Phantom.js and SharpKit

Dmitri — Fri, 02 Dec 2011 15:27:39 -0000

No, that would just change 'ize' to 'ise' and the like.

Re: Web Mining via Phantom.js and SharpKit

hodzanassredin — Fri, 02 Dec 2011 15:09:34 -0000

Читаю постоянно статьи на английском, но эта наиболее непонятна для меня. Британский вариант английского?

Re: Chained null checks and the Maybe monad

Guest — Tue, 25 Oct 2011 10:51:29 -0000

just ... why?!

Re: Event Brokers in .Net

Dmitri — Thu, 08 Sep 2011 16:15:20 -0000

Hi, I suggest you check out the example posted on BitBucket - it works just fine.

Re: Building a conversation engine with WebSharper

Dmitri — Fri, 11 Feb 2011 11:29:36 -0000

I no longer have it. But I'll be sure to post bugs now that I know where to post them.

Re: Building a conversation engine with WebSharper

Anton Tayanovskyy — Thu, 10 Feb 2011 07:04:58 -0000

I'd like to get a copy of the code that emits JS in the wrong order to fix it:

https://bugs.intellifactory.co...

Thanks,

Re: JSON Processing in F#

Larry — Fri, 28 Jan 2011 00:52:35 -0000

Really interested in how you recommend using F# with MongoDB (which has a recently released C# driver)

Re: Continuous Testing

Dmitri — Fri, 03 Dec 2010 11:40:42 -0000

I think the idea of things running as you finish writing a statement is very nice, but also unrealistic given how slow compilation actually is (even if you are compiling a Hello World project).