|
PureBytes Links
Trading Reference Links
|
NeoTicker can run with UMDS M3 as the backend for a long time :)
Obviously no one notice that :)
About using VB ... it could work for probably personal projects
but for full market scanning and tick by tick processing, it is
totally infeasible to use VB as of the current CPU speed and VB speed.
Maybe that will change when P5-10G is available :)
The experience from my programming friends in high-end vertical
markets on farm out of programming work is scary - each piece of code
seems to work, but the outputs at times are "typed-in" hoax and that
quality control becomes nightmare ...
Plus there is no real specs to a growing project - clients always
want more features after the current version is fully exploited.
Thus, in-house experts are better in the case of vertical platforms
like trading tools when consistent changes and improvements are required.
Cost - at a million line of debugged quality C++ code, the going rate
is at least $1 per line excluding cost of UI designs (very expensive),
plus documentation, etc. The rate is not cheaper at oversea countries.
The most expensive and under estimated cost in the inet boom and bust
is in fact this "development" cost :)
-Lawrence
--- Thomas Alexander <alexander_enterprises@xxxxxxxxx> wrote:
> With all this discussion about a TS replacement, I have not heard anything
> about using the UMDS (Universal Market Data Server)product as the back end. In
> my software company days we were planning on using it with a VB front end. The
> product seemed realiable, very fast, good VB and C++ interfaces and cheap! It
> even came with a simple charting package and accepted data from a number of
> major data sources. They also had licensing rates, etc.
>
> IMHO: As of a year and a half ago this seemed to be the best infrastructure
> upon which to build a TS for the new millenium. Any tech heads have any
> comments on this or looking into this avenue? Of all the coders I've hired, VB
> guys are the cheapest market rates and in this high tech economy... As a matter
> of fact, given a good spec someone could parse the work out to Indian firms at
> half the US rate! A good spec should be easier also as the product has already
> been well defined over the last decade, this is not a new thing being done.
> I've done a lot of software development in the last decade and I would be
> suprised if an operation run by experienced software development management
> could not do it in less than a year for about $100-$200k. Even less if you know
> how to negotiate employee option plans... And no, I won't be :)
>
> Just a few high altitude thoughts.
>
>
>
=====
Lawrence Chan http://www.tickquest.com
Transform market data into opportunities
|