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

RE: System tester overhaul (was Re: Question on calculation)



PureBytes Links

Trading Reference Links


Sure would be easier if your formula language was object oriented.

Gary

-----Original Message-----
From: PD Manager [mailto:pdmanager@xxxxxxxxxxxxxxxxx]
Sent: Monday, July 24, 2000 11:15 AM
To: 'metastock@xxxxxxxxxxxxx'
Subject: RE: System tester overhaul (was Re: Question on calculation)


Wherever the information regarding MetaStock coding practices came from,
that information is incorrect.  MetaStock design and coding methods are
not
from the 60's or 70's.  MetaStock is NOT a single, monolithic, top-down,
procedure oriented mass of spaghetti code.

MetaStock has been coded using methods of the 80's and 90's.
Event-driven
programming and object oriented design and coding have been used.  Now
to
deflect all the comments that I am sure are coming:  Obviously this
hasn't
helped us produce bug-free software. Regardless of which methods are
used,
we are trying to produce quality software that you, our customers find
useful at a reasonable price.

For the record, most of the difficulties involved with adding new
features
to MetaStock (or any program) are caused by trying to maintain
compatibility
with previous versions of the program.  Baggage from the past tends to
make
new features difficult to implement.  I submit the latest round of
problems
with upgrading custom indicators etc.  We have tried to maintain a
capability to upgrade all formulas back through version 2.x.  

At some point we will have to abandon backward compatibility in order to
produce the capabilities that our users want.  At that time, we will
have to
take our lumps because we have made some older versions obsolete.

As stated before, I am not here to do a public relations job, nor am I
here
to defend Equis or make excuses.  I am here to clear up misunderstanding
or
misinformation when necessary.

Ken Hunt
Programming Manager
Equis International



-----Original Message-----
From: Lionel Issen [mailto:lissen@xxxxxxxxx]
Sent: Saturday, July 22, 2000 7:18 AM
To: metastock@xxxxxxxxxxxxx
Cc: PD Manager
Subject: Re: System tester overhaul (was Re: Question on calculation)


Think it was sometime last year, in response to requests concerning V
6.5x,
Equis said that one of the big difficulties in making changes was that
Metastock was written as one large program.  This means that the program
is
written using the programming methodologies of the early main frame
computers circa 1960's.

Back around 1970 IBM developed a modular programming methodology with
the
objective of avoiding these problems. IBM published the details and made
presentations of this methodology at relevant meetings. I think that it
was
called structured programming or top down programming.

About 2 years ago, Scientific American had a feature article on error
free
programming. The article referred to work done in France and in the US.
The
article mentioned that most software companies won't use it because: "it
is
too expensive" and "it takes too long". Actual investigations showed
that
the overall programming costs are cheaper because there are fewer
revision
and bug removal costs, and overall programming time is also much less.

Lionel Issen
lissen@xxxxxxxxx
----- Original Message -----
From: Owen Davies <owen@xxxxxxxxxxxxx>
To: <metastock@xxxxxxxxxxxxx>
Sent: Friday, July 21, 2000 10:41 AM
Subject: Re: System tester overhaul (was Re: Question on calculation)


> Among other useful suggestions, Ron Stockstill requested:
>
> > 1.  Use metastock formulas for calculating the actual entry price
> >     and the exit price.
> >
> > 2.  Provide functions that will return the entry price and exit
price.
>
> Yes, yes, yes!
>
> Also, though I know nobody else probably much cares, I would
> really like to be able to see my system results in dollars, rather
> than points.  I'd much appreciate a database of futures contract
> details to use with system tests -- conveniently editable, of course,
> so we can update it after changes.  And the usual specifiable
> fudge factors for commissions and estimated slippage.
>
> But if you'll just give us the first two, I'll happily settle for it.
>
> Owen Davies
>