Re: [Chicken-hackers] CHICKEN in production

[Yaroslav Tsarko]

Hi,

On 07.10.2014 22:04, Peter Bex wrote:
> > There are hot internal discussions currently about migrating to something more 
> > widely supported (with proper debugger, profiler, and other useful tools) for 
> > our next big project, because a new hardware is more powerful and there are 
> > fewer restrictions.
> This is a bit of a tidal function: sure, hardware gets faster every year,
> but then they invent some new class of device which is more constrained
> than the previous generation, or a new niche of computation evolves where
> every CPU cycle is precious (bitcoin mining? 3D games?).  So even though
> there are lots of people falling over eachother trying to tell you that
> "hardware doesn't matter" and you should use their slow-ass language,
> that's just bullshit: performance will *always* matter.

My 5 cents. Oleg mentioned internal discussions about migration to 
something other than CHICKEN. To be more specific those discussions take 
place not because of CHICKEN performance issues (luckily, we hadn't too 
much of them) but because of lacking of development tools.
I mean, first of all, debuggers and performance analyzers/profilers. It 
is counter-productive to sit and peer at the code thinking why it 
doesn`t work. Yes, sometimes bugs are quite obvious, but more often they 
are not. So using debugger will increase speed of fixing bugs 
dramatically. Instead of debugger you have to insert additional logging 
which pollutes the code and makes you change its structure only because 
of adding logging (the most trivial example of this is when you want to 
look at function`s returning value - you have to wrap it in synthetic 
let only to capture its return value and log it).

Logging is also the only mean to perform performance measurements. And 
unlike log-driven "debugging", this is really slow and frustrating 
process - if you don`t sure where the bottleneck is, you spend a *lot* 
of time inserting logs here and there trying to catch the right point. 
Add to that recompile/relaunch cycle (chicken -> c -> gcc -> ldd -> 
switch to SSH console & launch of firmware -> replicating the 
performance issue) and you easily can spend a day trying to catch 
something not so obvious.

Some time ago someone in this (or maybe chicken-users) mail-group 
mentioned the same problems which led them to switch from CHICKEN to 
Python. This time situation is similar, although final decision about 
transition is not made yet.

--
Thanks,
Yaroslav