|
ECNs, Exchanges
< Previous Next >
Re: Dark/Lit Indication on Execution Reports
Hanno Klein / Deutsche Börse Systems <>
18 Jun 2010 5:24AM ET It was my personal opinion as there is no general statement from FPL that anything that is not defined in the spec as permissible behaviour is deemed a mis-use. Posts in the forum must be seen as personal opinions if they are not backed by the spec or an official Recommended Practices document.
FPL has intentionally chosen not to require certification of each and every interface spec out there. The downside is that often enough, many different interpretations exist for the same concept. The reason may be as simple as not wanting to go through FPL and ask for a FIX extension and to just look for the best field(s) that are not limited to a list of pre-defined values and define one's own stuff in there.
I will use this post to encourage people once again to come forward and work with FPL to make sure their usage of FIX aligns with the intentions of the spec and to put in the extra effort to go through a Gap Analysis process to get extensions where they are needed. It is less work than people might think and it saves the community a lot more work down the road when they do not hit all these different interpretations and have to code to them.
Part of the beauty of FIX is its flexibility. Lack of enforcement should not encourage mis-use as it is a disservice to the FIX community and requires money that is better spent elsewhere. Mis-use sounds like it is always intentional which it is not. Lack of knowledge regarding FIX concepts is a real issue. Part of the solution is to refrain from only reading message layouts and not the surrounding text of the spec. A little more time spent upfront before designing a FIX interface is a good investment for the benefit of the entire FIX community.
> Yes. That is the intension that is apparent from the specification. However, I recall seeing posts on this discussion forum about also using TradingSessionID (336) to denote separate books (e.g. odd lots). I was curious whether such a use would be deemed appropriate.
>
> > TradingSession(Sub)ID has a different intention, i.e. to represent (sub-)phases in terms of time, not location. FIX 5 SP1 defines a list of recommended valid values even if you can still use any value for backward compatibility.
> >
> > TradingSessionID is for larger phases that can be expressed in terms of the business day (e.g. 1=Day, 2=Half Day,...). TradingSessionSubID is for shorter phases that are part of TradingSessionID (e.g. 1=Pre-Trading, 2=Opening,...).
> >
> > > Would TradingSessionID (336) be another option?
> > >
> > > > I am not aware of a standard here but I am not sure distinct MICs are the best approach. You do not want a hierarchy of MICs, one for ExDestination and sub-MICs for the lit and dark book. Also without a hierarchy, there is often a lot more attached to a MIC, e.g. trading admissions which you do not want to duplicate. Something more lightweight is needed here.
> > > >
> > > > There is a user defined field (5018 OrderBookID) that goes more in the direction I think we should be going. Another option is to extend an existing tag (1021 MDBookType) beyond market data. It covers all kinds of situations where you have distributed liquidity of a single instrument within the same trading venue, i.e. round lot vs odd lot book, retail vs institutional order book.
> > > >
> > > > > > Is there a (burgeoning) standard mechanism for indicating that an execution has occurred on a dark or lit segment of an integrated book?
> > > > >
> > > > > No market standard. Using FIX 4.2 some firms will modify tag 30 on exec reports to distinguish.
> > > > >
> > > > > Personally I think we should see 2 ISO MICS - light and dark - per ex destination....
Re: Dark/Lit Indication on Execution Reports
Hanno Klein / Deutsche Börse Systems 18 Jun 2010 5:24AM ET
|
