@maleghast I love First Direct. Nothing but good from them. Malcolm will say the same too.

http://www.brics.dk/automaton/index.html this is really fast. It outperforms java regexes by quite a bit. Has a fairly functional api too. I'm experimenting with building a router where the goal is "fast" first, getting some interesting results.

@dominicm sounds cool!

@dominicm - Glad to hear that things have improved with them, or at least that they manage to give good service to some people…

Looking at the link re regexes

1 gotcha: no capture groups.

Yes, I just noticed that in the FAQ

I imagine that there is value in using it where you need speed over flexibility, but reverting back to more standard pattern matching approaches for other tasks..?

I was hoping I could emulate it via info exposed from the engine. But I'm reaching my giving up point.

You were going to emulate it to re-do the capability in Clojure?

Or you were going to emulate capture groups?

emulate capture groups. Routes only need a few simple ones.

Just for comparison, for static routes, I get numbers like: brics: 6ms, regex: 15ms, silk: 50ms, bidi: 115ms.

It's probably worth using brics to filter down the route, then use regex to pull out the groups. If you have ~30 routes, it's 6*30 + 15 vs 15*30.

Brics is advantageous that you can build it up with functions, unlike regex which is careful string concatenation at best, which really sucks.

String concatenation just is rubbish.

Ah I see, you are optimising routing…

*light dawns finally on maleghast’s mind*

I have spent the morning reading “scholarly” works about Supply Chain Risk Management and realising three things: 1. There are almost no Open Source tools in this area 2. Almost all of the theory springs out of Change Management theories and “frameworks” 3. If I could just get my head around it I think that logic programming is going to be a huge help in trying to do this stuff…

My brain is now melting.

I am also secretly of the opinion that a LOT of the “research” and “scholarly work” in this area is a lot of hot air and pontificating and not a lot of ACTUAL work, and that Management Consultants and McKinsey style “trouble shooters” have been making a LOT of money out of spouting off about this stuff and not actually having to deliver anything.

What is Supply Chain Risk Management exactly? It sounds quite customer-specific.

Basically it’s about modelling the risks inherent in a supply chain, and then having predictors and monitors to empower decision making… Imagine you are Unilever, and you have to maintain clear routes of supply for staple ingredients for your food products, like soya, sugar, wheat etc. The business process of modelling the risks inherent in that supply chain extend to geophysical risks (weather, climate, geography and geographically specific disasters like flooding landslides etc.), political (taxation, trade tariffs, even political instabilities) and logistical risks (such as reliability of shipping, time-based risks represented by port activity, etc.)

It is, to an extent, something that cannot be genericised, but there are examples of people trying to provide generalised tools like this: https://www.riskscape.org.nz/ and I am looking into ways we might be able to do it for agriculture and specifically supply chain risk instead of just geophysical risk (like RiskScape)

and it’s melting my brain…

But I’ll get there… There is this thing called Discrete Event Simulation that might be well suited to Logic Programming and / or DAG stuff…

I need to do some more thinking…

Now, lunch time…

Right, gotta go get ready to leave for the airport - London I will be in you later… 🙂

@maleghast I don't do that until 0dark30 tomorrow

Night flight or the night train, @otfrom ?

Morning flight