On 05 Mar 2010, at 20:14 , Andrey Fedorov wrote: > First, let me apologize if I came off as rude before - I didn't mean to, was > just in an aggressive mood.
I get moody too. For instance I have no idea why I'm having this conversation - when I got out of bed this morning I could have sworn I wanted to spend the day coding. %-} > Cool! I'd never heard of CRT before. Thanks for explaining it. Just to make > sure I understand, it's a two-color directed graph (the "bananas" are one > color, the "normal" nodes are another), with the restriction that each banana > node must have exactly one outgoing edge into a normal node? CRT's are not directed graphs. Directed graphs do not define semantics. You are correct though in that you can describe the topology of a CRT as a two-color directed graph and, in fact, that's exactly how these folk represent it: http://flyinglogic.com/ > I have found it to be a powerful tool for modeling program execution. > > It's certainly a neat way to look at large programs. I'm not quite sure how > accurate it could be at describing biological systems or the like. As Alejandro pointed out, CRT modeling requires that we have at least an idea of the predictable processes in the system! Therefore the limiting factor on the accuracy with which you could use a CRT to describe a biological system is the degree to which you have a grip on the predictable processes within the biological system. So, for instance, something like ATP - ADP cycle can be modeled quite accurately and with considerably less energy requirements than a full-blown molecular dynamics simulation of the offending proteins. But protein folding? Not so much no :-) > On Thu, Mar 4, 2010 at 8:20 PM, Alejandro Garcia <[email protected]> wrote: > On Thu, Mar 4, 2010 at 3:09 PM, Andrey Fedorov <[email protected]> wrote: > The picture you gave isn't a system, it's a directed graph. I guess you're > implying anything you imagine to be a "system" can be represented as a graph > - but what is a system? > > Well it isn't a system in the same sense that a map isn't the terrain. I > think people call those things a representation. > > Precisely! So we *are* on the same page. It's a representation which doesn't > always preserve a system's "complexity" Careful here, the particular use of the word complexity here is different to the one you're using: i.e. It's used in the sense of: An orbital simulation does not preserve all the details of the planets (what color are they, are they made of cheese?), but it does preserve all the details of how they interact with one another. So, this is complexity in the sense of: #'s of independent processes and their degrees-of-freedom in respect of one another > (without defining "complexity"). So all I'm getting from your earlier point > is that the CRT representation of a system can't be used to define > "complexity". So it's a crappy representation, after all. Not to 'define' complexity no - but to show one kind of complexity in terms of showing all the 'moving parts' within the system and how they connect to one another yes. > If we *do* want to define "complexity", we could put a constraint on these > CRT graphs, like "nodes have no state"? 99% of the effort in preparing a well-factored or 'dry' CRT on a particular subject is becoming conscious of how processes you once viewed as really complex and mysteriously stateful actually decompose into a handful of relatively simple interactions. It can be a highly emotional experience. > This is starting to smell like the classical argument against OOP. If you were fortunate enough to enjoy a school or teacher that encouraged contemplation of the word "encapsulation" then it's pretty much down to six of one, half a dozen of the other and precisely how much bondage and discipline you like from your compiler IMNESHO. But I'm going to STFU now because the EarningOfTheLivingWage(tm) doesn't leave much time for the pursuit of my own projects so if I ever hope to make any lasting contribution to my world I'd do well to start asking myself where my efforts are best served. - antoine _______________________________________________ fonc mailing list [email protected] http://vpri.org/mailman/listinfo/fonc
