Re: Corollas, Not Maseratis (Re: Whose Fish?)


S Perryman wrote:
topmind wrote:

S Perryman wrote:

TM>Then what the hell is the *claim* by Mr. May etc?

SP>The statement that the financial services sector is a very large domain,

So are a lot of others: transportion (airline, railroad), law
enforcement, etc.

Actually, I did some work for a large cable company, which is
technically a "telecom" (they offered phone also, although
subcontracted some of it), and a lot of their stuff was rank-and-file
biz apps. I was not aware of *any* graph optimization/traversal
systems used there. (I'm sure they existed, but were certainly not
rampant. They may have rented existing software for that). I also did
a breif stint at a phone-only telecom, and never saw graph opt/trv
there either.

Thus, my experience seems to contradict that statement.


TM>I won't necessarily challenge that, other than saying I suspect a lot
TM>of it is just rank-and-file custom biz apps outside of the
TM>optimization of trade order or selection kind of stuff. But even if
TM>true, it is not material to our debate.

SP>and that the choice of paradigm there is important.

TM>Not disputed, or at least not being challenged by me.

SP>And that OOP has been
SP>chosen as a paradigm for use in that domain.

TM>Not proven.

Proven.
Used (certainly in Europe) by financial institutes for a wide
range of problems (share trading, futures, equities, derivatives etc) .

"Proven" would involve figures along the line of:

Survey Organization: XXXXXXXXXXXXXXXXXXXX
Managers surveyed: 9999
Developers surveyed: 9999
Companies surveyed: 9999
Mangers Who Prefer OOP: 99.99%
Developers Who Prefer OOP: 99.99%

1. No it would not.

The latter two would have "Prefer OOP" replaced with 'think OOP
has helped develop the same/better as legacy approaches" . Something
is proven when it X does the job either better than or as well as Y, not
when someone states that they prefer X over Y.

Okay, I'll accept that correction. However "equal to" doesn't count.
Personal preference trumps "equal to". Also, if we are going down this
nitty road (the above was meant as a rough draft to convey the idea),
then the percentage of their apps that are OO may also be an issue.
See below.



2. And "not proven" would involve figures along the line of:

Survey Organization: XXXXXXXXXXXXXXXXXXXX
Managers surveyed: 9999
Developers surveyed: 9999
Companies surveyed: 9999
Mangers Who Do Not Prefer OOP: 0.01%
Developers Who Do Not Prefer OOP: 0.01%

So what figures do you have for your claim of "Not proven"
(rhetorical question) ... ??


And if it is the most common, you still have not shown

whether it was by choice or fad chasing.

Well when Nomura (the HOODINI project) and the Swiss Bank in London
decided to use OO, they did so in the context of the legacy development
envs they already were using. Given the scale and costs (in development and
potential loss of business/revenue, the decision was not made lightly.

That says nothing about the companies that didn't use OO. One company
is not a survey.

1. Two companies. Two very very big banks in London who were subjected to
quite a lot of industry s.

2. Which companies are not using OOP in the financial services sector ??

Using a criteria you were quick to cling to (jobs posted in a 24 hr period
for banking work) , a very low percentage it would appear.

It is impossible to avoid *some* OO these days because standard tools
and API's use it. It is the QWERTY paradigm. Thus, a thorough survey
should also include the percentage of OO. Perhaps a distinction should
also be made between creating classes and using classes. Trickier
still, many put stuff in classes that don't belong there: they are
merely procedural modules in OOP wrappers. I find it hard to call an
app that does not use polymorphim or inheritence "OOP".



TM>Do you think Windows is the
TM>most popular OS because it is technically the best? A survey would be
TM>needed to settle this.

Fortunately, we have people here such as Patrick May who have used OOP
in this domain. They may tell us something about the inherent nature of
the domain that makes OOP worth consideration.

Don't tell, show. We need yet more anecdotes like a hole in the head.

Well, he did "show" .
A trading settlement system that used graph-theoretic techniques.
Perhaps he will "show" us some other problems and the systems that
solved those problems.

Are you talking about that one that Freeb produced an initial
solution, and then P. May introduced yet more specifics *after* he saw
Freeb's draft? I can understand why Freeb abandonned that one.

I thought you claimed somewhere that OO was not really better for
graph traversal/optmz anyhow. This seems to be a contradiction of your
stance. Please clarify.



Regards,
Steven Perryman

-T-

.

Relevant Pages

  • Re: Corollas, Not Maseratis (Re: Whose Fish?)
    ... Mangers Who Prefer OOP: 99.99% ... Developers Who Prefer OOP: 99.99% ... Survey Organization: XXXXXXXXXXXXXXXXXXXX ... Which companies are not using OOP in the financial services sector ?? ...
    (comp.object)
  • Re: Corollas, Not Maseratis (Re: Whose Fish?)
    ... Survey Organization: XXXXXXXXXXXXXXXXXXXX ... Developers Who Prefer OOP: 99.99% ... Steven Perryman ...
    (comp.object)
  • Re: Understanding Public Module or Class
    ... samples to the OOP, if I had seen that it was in this newsgroup I had made ... VB.Net makes it too easy for former classic VB developers to migrate to ... The .Net platform and the Common Language Infrastructure are truly ... object-orientation. ...
    (microsoft.public.dotnet.general)
  • INTERVIEW: Why do people contribute code for free?
    ... The title of the survey is FLOSS-US which stands for Free/Libre Open Source ... monetary and non-monetary -- that drive the developers. ... expectations that the developers have from making these contributions, ...
    (comp.os.linux.announce)
  • Re: Controls v HTML
    ... To think that OOP is something new to Classic ASP developers is not so, although not all asp developers may take such a approach. ... First, you don't seem to understand the difference between pseudo-object-oriented programming technology and true object-oriented programming technology, and the difference is significant. ... It would seem to be new to you, but OOP is been around for longer than Dot.Net and is not the relient on using controls as you seem to state in your first post. ...
    (microsoft.public.dotnet.framework.aspnet)