PureBytes Links
Trading Reference Links
|
Vitaly,
I'm still waiting for v7.02. I just thought I'd pass on kudos for v7.01. I
couldn't believe the speed improvement.
Guy
Paranoia...you only have to be right once to make it all worthwhile!
-----Original Message-----
From: owner-metastock@xxxxxxxxxxxxx [mailto:owner-metastock@xxxxxxxxxxxxx]On
Behalf Of Vitaly Larichev
Sent: Friday, July 14, 2000 10:04 AM
To: metastock@xxxxxxxxxxxxx
Subject: Re: What are the FIXES in 7.02
Guy,
Much faster performance with v. 7.02 is really great news! Thanks Equis!
I used to complain about MS's lackluster job in previous versions. One of
the
reasons was an unfortunate way chosen to code "IF" logic: it would calculate
ALL expressions before checking the branching condition, i.e. in IF(A>0,B,C)
it calculates A,B,C before checking if A>0, although only one of B and C
values is really needed due to A>0 condition. In the following example
If(2>1,1, 1/(1-1))
it would give an error trying without need to calculate 1/(1-1) . If the
change
you noticed was also due to corrected "IF" logic, it would save us a lot
of time in explorations, etc. by nesting "IF" conditions in such a way that
the first one, fast and easy, would eliminate a majority of candidates
whereas
others, more time consuming would deal with much smaller number of stocks,
sort of sequential approach.
So, it's a long preface to ask you make an indicator with the formula above
and see
if it doesn't anymore give the "division by zero" error in v. 7.02 as it
does in
my 6.52. It would take a few seconds of your time while would be of great
help
to me to decide on upgrading.
Thanks a lot.
Cheers, Vitaly
Guy Tann wrote:
> Ken,
>
> I thought I'd send you guys a compliment about 7.01 after all the
brickbats
> you've received here.
>
> I finally got around to installing it yesterday (I've probably had it for
a
> few months) and the install went smoothly. Take that to mean no problems.
> It also converted our own indicators without missing a beat, unlike a few
> versions ago that ate them for lunch during the conversion (necessitating
a
> complete reprogramming effort and about 2 months worth of work to
completely
> retest everything and run parallel for a couple of months more).
>
> Since we don't do any charting, I was more interested in the calculation
> speed of the new version. Version 6.5.2 solved a few problems for us that
> prevented us from using MS ever since our DOS conversion. In fact, we
still
> used the DOS version since all prior Windows versions couldn't handle our
> calculations. In addition to poor memory utilization, the DOS version was
> able to use a virtual disk whereas the Windows version used our hard drive
> like a thrashing machine. I probably exchanged over 100 e-mails with tech
> support before giving up. I did purchase every upgrade and went through
> this conversion too many times to count.
>
> I'm happy to say that v7.01 version blows v6.5.2 away. Calculations that
> used to take almost a minute in v6.5.2 are done in 6-8 seconds in v7.01.
> For our work, that's absolutely fantastic. I'm starting to get excited
> here, as we might be able to use System Tester for the first time since
the
> old DOS days. We have already replaced our old DOS version of MS as well
as
> TAS. If this final conversion works, I'll be able to put my Clipper
> programs to bed as well and we'll be running under one system. After
> umpteen years with Equis, we finally bought a second package for my
brother
> since we were finally able to use it in a production mode.
>
> Now that we seem to have arrived at some level of stability, I plan on
> moving our Intermediate Term System into MS. I'll probably need to ask a
> few questions about variable handling and the previous. With a little bit
> of luck, I might be able to move my final DOS programs into MS and finally
> settle on one system and one database.
>
> Now, after saying all of those nice things, I looked up the previous
> function in the index and it pointed me to page 258. There is no previous
> function on page 258, or 259 for that matter. Did I get a bad manual or
is
> this a misprint? If you could, I'd appreciate your pointing me in the
> direction of the previous function in the manual.
>
> Thanks again,
>
> Guy
>
> Paranoia...you only have to be right once to make it all worthwhile!
>
|