Marketing of D - article topic ideas?

[Andrei Alexandrescu]

On 06/03/2010 08:34 PM, Adam Ruppe wrote:
> I recently rewrote a huge PHP application in D2. It isn't a completely
> fair comparison, since the PHP was written by, well to be polite, PHP
> "programmers", but the results were pretty astounding.
>
> 90% reduction in lines of code [!], 200% improvement in speed under
> heavy load (only 50% improvement under light load, still good, but not
> as good), and the code is actually maintainable! Also, the original
> took about a year for the first team to write. My D rewrite took 6
> weeks. Oh yeah, and the PHP is riddled with security holes. Not so
> much with the D.
>
> The best part about the speed is that it is just running as plain old
> CGI and my code does a lot of inefficient looping through XML. I could
> probably double the speed again by improving that.
>
>
> I might be able to write this into an article, but what I've done so
> far is just showed it off to other coders in some private messages on
> forums. A complication in publically showing it off is it is closed
> source though. The speed advantage isn't apparent until the code
> actually does something; PHP is much faster for hello world, but gets
> spanked when it comes to the actual work.

I think this is the best way to approach writing, and I strongly 
encourage you to just plow through whatever IP issues there are and go 
for it. (In my experience, nobody will come after you for writing an 
article that doesn't outright give the code away.) I'm repeating myself, 
but writing a good piece should at best start with a desire to share 
something you genuinely believe is interesting (as opposed to starting 
with a desire for developing an article and consequently looking for a 
topic idea). You seem to be right on the money.

In related news... For what it's worth, I plan to write about 
memory-isolated containers (those that can use malloc/free for their 
allocation strategy without being unsafe). I believe there are a number 
of fine points to be made and that such containers solve an important 
class of problems. InformIT.com considers the idea interesting and 
commissioned me to write that article.

I've also read again today this piece by Oleg Kiselyov:

http://okmij.org/ftp/papers/LL3-collections-enumerators.txt

Oleg is pretty awesome and makes great points, but I think also missed a 
few. The abstract is a bit difficult to get into (and I haven't 
understood his enumerator inversion trick without using continuations). 
I have a pretty good retort in my mind, but that would take quite a long 
time to write.


Andrei