C[omp]ute

Welcome to my blog, which was once a mailing list of the same name and is still generated by mail. Please reply via the "comment" links.

Always interested in offers/projects/new ideas. Eclectic experience in fields like: numerical computing; Python web; Java enterprise; functional languages; GPGPU; SQL databases; etc. Based in Santiago, Chile; telecommute worldwide. CV; email.

Personal Projects

Choochoo Training Diary

Last 100 entries

LMAX Architecture

From: andrew cooke <andrew@...>

Date: Sun, 30 Oct 2011 08:05:20 -0300

I'm spending much of this weekend trying to work out why I am sometimes
frustrated with work.  One - perhaps rather mundane and obvious - factor is
trying to balance conflicting and sometimes hard-to-understand requirements.
So reading an article like this, which describes a beautifully clean design,
driven by one simple objective - speed - is like reading a fairy story,
describing an idyllic land, where targets are clear and succeess easy to
measure....

http://martinfowler.com/articles/lmax.html

Andrew

too fast already

From: Stefan Wehner <stefan@...>

Date: Sun, 30 Oct 2011 10:18:00 -0400

Hi Andrew,=20

While I share your fascination about the clean, smart architecture and =
code, I cannot help to point out what the purpose of the system is: real =
time trading of derivative financial products. Sounds to me like putting =
together the nuclear bomb for financial markets. And it's a dirty bomb. =
Fowler in his footnote [2] comments about derivatives: "I shall remain =
silent on what I think about the value of this innovation".=20
The role of the LMAX platform, as I understand it, is to give more =
people (read: highly specialized traders) access to high speed trading. =
High speed involves that the actual trading is done by algorithms. The =
trader defines the algorithm, but the machine runs it. If there were =
still people taking the buy-sell decision, you would not have to be as =
fast as 6 million orders per second. This is far beyond human speed.=20
My conclusion is that the mess will make its appearance eventually. You =
are right, you don't see the mess in their code and specs. But that's =
exactly because they completely disregard the effects their trading will =
have on the real world. I would say, their spec is just incomplete! In =
the end, this kind of innovation will cause unstable financial markets, =
an unstable economy, financial crises, concentration of the wealth in =
the hands of very few, people losing their jobs and their houses, =
radicalization of politics, kids without a future (I have many of them =
in my neighborhood in NY). The people who will later attempt to clean up =
the mess are very low payed or volunteers, social workers, underfunded =
government agencies etc.=20
Of course, this is not the fault of LMAX all this has been going on for =
a long time. I believe we rather need to think about how we can partly =
slow down, rather than further speed up our life. There are too many =
side effects we don't see before it is too late. And there are too many =
already who cannot keep up with the speed of this world.

Saludos,
Stefan

Re: too fast already

From: andrew cooke <andrew@...>

Date: Mon, 31 Oct 2011 17:33:53 -0300

I don't disagree with you about high frequency financial trading - I was more
interested in the technical details than the application.

However, I do wonder if there is a correlation between evil and simplicity.
Focussing on one aspect of a problem to the exclusion of all else leads to
simplicity, but imbalance.  So perhaps it is in the nature of things that the
cleanest technical problems are also the most damaging...

Andrew

Comment on this post