[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Bad Initial Ticks



PureBytes Links

Trading Reference Links

Janette Perez wrote:
 
> The TS 4.0 server is architected to store daily ticks as an offset of
> the first days trade. As a result, if the first tick transmitted by the
> data feed is off, the first tick of every data block during that day
> must be manually edited. While, thankfully, a bad initial tick isn't the
> norm, when it does occur we realize it requires a great amount of time
> and inconvenience to correct. We sincerely apologize for this fact.

Step 1: Admit TS 4 has an insidious problem.

> 
> I can assure you, however, that one of our top objectives in designing
> our next generation server and applications has been to take these kinds
> of design decisions into account to insure, among other things, that
> data feed errors on the first tick of the day do not cause anywhere near
> the kind of inconvenience they cause today.

Step 2: Put off doing anything about it, but do it is such a way
as to ignore the customer's direct request.

> 
> Thank you very much for your continued support, patience and input. All
> three are truly valued and appreciated.

Step 3: Ask for forgiveness of your sins, both in the past and 
the new one committed by your non-answer.


Excuse me, but your note is one of the best worded, non-answers
I have ever received from an ISV (independent software vendor).
Not only did you completely ignore my request for a fix to the bad
initial tick problem in TS version 4, but you conveniently did not
even say if it will be fixed in the first release of TS version 5.
All you could say was you have tried to design the next generation
to eliminate this and similar inconveniences.  Not a word of 
assurance that your design does indeed eliminate this problem.
Not a word of assurance that the existing implementation of your
design does indeed eliminate the bad initial tick problem.  

Since beta testing is in progress, surely somewhere in your office
there was a TS 5 system collecting data for MRK from a BMI feed
this week.  Is the problem fixed, or not, in the TS version 5 
currently being tested?

Lastly, you need to fix this problem in TS version 4.  Directly 
address this request, please.

Rod Grisham