Dolphin 6.0

When should we expect Dolphin 6.0 ?

> When should we expect Dolphin 6.0 ?

I should expect I will be the only reply to this message ;)

Soon. Soon.

Sean Malloy wrote:
>>When should we expect Dolphin 6.0 ?
>
>
> I should expect I will be the only reply to this message ;)
>
> Soon. Soon.


Well, what does soon mean?  I was planning on switching from VAST to
Dolphin for development but if soon "means a year", it's not soon enough
for me...  And I don't want to buy version 5.x if they'll release 6.0 a
few months after...

> Well, what does soon mean?  I was planning on switching from VAST to
> Dolphin for development but if soon "means a year", it's not soon enough
> for me...  And I don't want to buy version 5.x if they'll release 6.0 a
> few months after...

No idea. There was talk of "Soon" over a year ago.

Anyways, I had the same objection to buying dolphin XP when I knew 6 was
going to be out, Andy from Object Arts assured me that to upgrade to Dolphin
6 professional, I would only have to pay the difference (If there is one)

I haven't heard anything to challenge that, so I assume it still stands.

Anonymous,

> Well, what does soon mean?  I was planning on switching from VAST to
> Dolphin for development but if soon "means a year", it's not soon
> enough for me...  And I don't want to buy version 5.x if they'll
> release 6.0 a few months after...

I've included a post from Andy Bower on the 16th Feb that should allay any
cost fears about buying D5 (and may help answer the when question) ....

=_=_=_=_=_=

Folks,

We are receiving a fair number of e-mail enquiries regarding the
upgrade policy for Dolphin 6 when it becomes available.

While we have not currently set the pricing for the new product (old
hands will know that we rarely do this until just before release when
we scribble a few numbers on the back of an envelope) it is likely that
the release will be sometime after the end of May this year.

In the meantime we have decided that all Dolphin 5 users who have
purchased the full product or upgrade after the start of 2004 will be
afforded the opportunity to upgrade to D6 at only the difference in
price. Hence if you purchase D5 Pro now at $349.95 and the equivalent
D6 product is $399.95 (remember, the new prices have not yet been set)
then you will only need to pay the $50 difference.

Hopefully this will address the concerns of those people who are
considering purchasing Dolphin right now but are wondering whether it
might be better to wait for the release of D6 before doing so.

Best regards,

Andy Bower
Dolphin Support
www.object-arts.com

=_=_=_=_=_=

--
Ian

Use the Reply-To address to contact me.
Mail sent to the From address is ignored.

"Anonymous" <[hidden email]> wrote in message
news:[hidden email]...
> Sean Malloy wrote:
> >>When should we expect Dolphin 6.0 ?
> > I should expect I will be the only reply to this message ;)
> > Soon. Soon.
>
> Well, what does soon mean?  I was planning on switching from VAST to
> Dolphin for development but if soon "means a year", it's not soon enough
> for me...  And I don't want to buy version 5.x if they'll release 6.0 a
> few months after...

I am guessing that we might see Dolphin 6 within the next few months.  I
suspect that most of us Dolphin 5 users are eager to get Dolphin 6 not
because of any serious flaws in Dolphin 5, but mostly because we want a new
toy to play with. ;)  Dolphin version 5 is a great system.  I switched from
VAST 4.x to Dolphin 3.x years ago, and I am glad I did.  Dolphin works
better for my needs (Windows PC based software) than IBM VAST did, and is
much more economical.

I just would not want people to assume that there is anything wrong with
version 5 based on our eagerness for version 6.  I am looking forward to
Dolphin 6, but I am still enjoying version 5. ;)

Chris

Folks,
 
Well, in a unprecedented display of openness ;-) let me try and explain
the current position regarding Dolphin 6. As you may have guessed (by
virtue of the fact that D6 is not already out there) we are running
late with the development. There are, in the main, two reasons for this:

1) Whilst the feature set for D6 is steadily growing, there are a
number of features that we had hoped to include that are not there yet.
We feel that it is worth delaying the release a short while to include
at least some of these. I don't want to print the shortlist here,
however, just in case your particular favourite doesn't make "the cut".

2) More importantly from the delay perspective, over the last two
months we have been engaged in some development that is aimed at
putting the company on a more secure financial footing. I can't really
disclose the full details of this but the work is not consultancy and
the end result will hopefully be a "novel" way of securing Object Arts'
stability into the future. We hope to have this effort complete (at
least to the point where it will only require minimal ongoing
maintenance) by the end of this month (May 04). After this point we
will be able to get back to full strength on Dolphin 6 development.

I apologize to all our users who are eager to get hold of the next
major version of Dolphin but, realistically, I would guess that these
delays will mean that you are unlikely to be able to do so before the
end of this year. Even so, I hope that most of you will agree with
Chris' comments above (that Dolphin 5 is a good system) and will be
able to weather this delay with only minimal inconvenience.

I would also like to re-iterate my previous comment about D6 upgrades;
Anyone who purchases Dolphin 5 from January 2004 onwards will be
upgraded to Dolphin 6 (when it becomes available) at only the
difference in price. Hence you won't be penalized by buying, or
upgrading to, Dolphin 5 now.


Best regards,


Andy Bower
Dolphin Support
www.object-arts.com

Andy,

> 2) More importantly from the delay perspective, over the last two
> months we have been engaged in some development that is aimed at
> putting the company on a more secure financial footing. I can't really
> disclose the full details of this but the work is not consultancy and
> the end result will hopefully be a "novel" way of securing Object Arts'
> stability into the future.

My intense curiosity aside<g>, this sounds great.  Good luck with it!


> I apologize to all our users who are eager to get hold of the next
> major version of Dolphin but, realistically, I would guess that these
> delays will mean that you are unlikely to be able to do so before the
> end of this year. Even so, I hope that most of you will agree with
> Chris' comments above (that Dolphin 5 is a good system) and will be
> able to weather this delay with only minimal inconvenience.

D5 is excellent, and I am in no particular bind.  As is usually the
case, I am more interested "not before" estimates (to not be caught by
surprise) than I am in precise delivery dates.

I would like to see you update the D5.1 installers.  The updated DLLs
are greatly appreciated, but require effort to place on "other machines".

Have a good one,

Bill

--
Wilhelm K. Schwab, Ph.D.
[hidden email]

"Andy Bower" <[hidden email]> wrote in message
news:[hidden email]...
> Well, in a unprecedented display of openness ;-) let me try and explain
> the current position regarding Dolphin 6. As you may have guessed (by
> virtue of the fact that D6 is not already out there) we are running
> late with the development. There are, in the main, two reasons for this:

I appreciate the communication and the situation.  If at some point you can
"spill the beans" about item two I am sure we would all be curious to hear
about it.

> I apologize to all our users who are eager to get hold of the next
> major version of Dolphin but, realistically, I would guess that these
> delays will mean that you are unlikely to be able to do so before the
> end of this year. Even so, I hope that most of you will agree with
> Chris' comments above (that Dolphin 5 is a good system) and will be
> able to weather this delay with only minimal inconvenience.

Thats OK, we can wait for our new toy. ;)  However if at some point there
were a beta available (even if not complete) I am sure quite a few of us
would be happy to kick it around for you, even under NDA if need be.  I
especially would like to encourage you to make a beta available to
third-party tool vendors before the official release to give them an
opportunity to migrate their tools as soon as possible.

Good luck with all your development.  We shall look forward to the results.

Chris

Andy Bower wrote:

[...]
>
> I would also like to re-iterate my previous comment about D6 upgrades;
> Anyone who purchases Dolphin 5 from January 2004 onwards will be
> upgraded to Dolphin 6 (when it becomes available) at only the
> difference in price. Hence you won't be penalized by buying, or
> upgrading to, Dolphin 5 now.

Andy,

Sorry, if I am asking what might be obvious, but does that also apply
to upgrading WITHIN Dolpin 5? In my case I have DVE5 and I am
currently contemplating to upgrade to the Professional Edition.

With Best Wishes for a speedy Dolphin 6.0 Development,

Bernhard

--
(Please remove all numerical digits from the
sender address for a valid email address. Thanks.)

Bernhard,

Yes. If you upgrade to D5 Pro now you will be eligible for D6 Pro (or
equivalent) at only the difference in the price between the full copies
of those versions. I'm sorry if this still sounds confusing but the
idea is simple; there should be no disadvantage for you to upgrade now
rather than wait for D6.

Best regards,

Andy Bower
Dolphin Support
www.object-arts.com

Andy,

Thank you so much for your confirmation, that was exactly what I was
looking for.

Bernhard
(now using D5 _Pro_)

Andy Bower wrote:
> Bernhard,
[...]
--
(Please remove all numerical digits from the
sender address for a valid email address. Thanks.)

Andy Bower wrote:
...
> I would also like to re-iterate my previous comment about D6 upgrades;
> Anyone who purchases Dolphin 5 from January 2004 onwards will be
> upgraded to Dolphin 6 (when it becomes available) at only the
> difference in price. Hence you won't be penalized by buying, or
> upgrading to, Dolphin 5 now.

This sounds great. My concern is backwards compatibility.
If I develop on 5, how likely is it that my application
will work out of the box on D6 too? I hope you take this
concern into account seriously.

D5 is promoted as the "XP" version. It would be great to
actually have a set of unit-tests within it, making it
sure everything works as before when stepping up to the
next version.

I also applaud you trying to get more financial stability.
Looking at http://object-arts.com/Home.htm it strikes me
that it states the site was last updated June 28, 2003,
almost a year ago. I understand there has been some patches
issued since, but there seems to be no mention of these.
I assume with more money, there'll be more marketing, and
a more up-to-date website too.

Thanks & best success
-Panu Viljamaa

"Panu" <panu@fcc.net_NOZPAM> wrote in message
news:[hidden email]...
> I assume with more money, there'll be more marketing, and
> a more up-to-date website too.

If we're lucky, we might even get documentation.

Panu,

I'm afraid we can't guarantee this. We usually try to maintain
backwards compatibility between "point" releases within a major version
(e.g. between 5.0 and 5.1) but we use the major version releases as an
opportunity to "burn the disk packs" when it becomes necessary. What I
mean by this is that, if there is some inconsistency in the Dolphin
class hierarchy which needs to be addressed or a new feature that
should be implemented but doing so would compromise compatibility, then
we like to have the freedom to proceed regardless.

That said, I can't think of any compatibility problems that D6 as it
stands would introduce for D5 users; it's just that we don't guarantee
compatiblity between major versions.

> D5 is promoted as the "XP" version. It would be great to
> actually have a set of unit-tests within it, making it
> sure everything works as before when stepping up to the
> next version.

Well, XP arrived rather late to the Dolphin development party. We do
have a growing number of unit tests for the base system but these are
being retrofitted as areas of the class library are revisited, either
for bug fixes or new developments. I believe this is the recommended XP
approach for integrating XP into an existing project (XP Explained,
Beck, p126). It's possible, I suppose that we could make these tests
available but there is far from complete coverage.

> I also applaud you trying to get more financial stability.
> Looking at http://object-arts.com/Home.htm it strikes me
> that it states the site was last updated June 28, 2003,
> almost a year ago. I understand there has been some patches
> issued since, but there seems to be no mention of these.
> I assume with more money, there'll be more marketing, and
> a more up-to-date website too.

There will be a new website and Wiki and online documentation for D6.

Best regards,

Andy Bower
Dolphin Support
www.object-arts.com

Chris,

As you may have seen on c.l.s, the beans are spilled here:

www.alchemetrics.org


Best regards,

Andy Bower
Dolphin Support
www.object-arts.com

Andy,

> As you may have seen on c.l.s, the beans are spilled here:
>
> www.alchemetrics.org

I hope this means that you have found something that is interesting in
itself, and will allow Dolphin to produce lots of money for you.


A few random comments/questions about the web site:


================================================
The have been 5 occasions (as of Oct 04) where an error in the Smalltalk
code has been detected at runtime (with a walkback) and has been fixed
by use of the built-in debugger.
================================================

That must mean that you elected (and there is obviously nothing stopping
me from following your lead, save an extra license or two) to run the
IDE on live machines.  Are there any pros or cons, other than the
obvious ability to fix errors on the fly?  Do you worry about recursive
walkbacks turning a small problem into a big one?



================================================
The Interactive Brokers Java client responsible for the broker API has
crashed 6 times causing several position entry problems. Indeed, the
issue was so severe that we now run the Java code on a separate computer
which can be rebooted when necessary independently of the other TWB
trading machines. Having said this, the IB client has become more stable
in recent months.
================================================

ROTFL.  This is almost exactly the experience I had with Novell client
software of years past.  They have improved a lot (not to be mistaken
for an endorsement<g>), but were at one time a BSOD waiting to happen.
We designed around it by the same technique.  I still tend to isolate
most "clients" this way, both because it reduces their ability to
clobber my machines, and because doing so often unloads a lot of work
from the main machine.


================================================
The Java client was auto-updated by Interactive Brokers on July 21st
2004 and that version refused to accept orders for the entire day
(insufficient testing?). The problem was fixed by another auto-update
the following day.
================================================

Ouch!  If you/they are doing what I assume you are, that's a lot of
money to have sitting idle (or worse trapped in a bad position) for 24
hours.

Dumb question: Interactive Brokers is a separate entity, that likes to
program in Java instead of sipping while working with a Smalltalk IDE?
Put another way, are you using Java because it was forced on you by
trading service?

Good luck!!!

Bill


--
Wilhelm K. Schwab, Ph.D.
[hidden email]

"Andy Bower" <[hidden email]> wrote in message
news:[hidden email]...
Very cool, thanks for sharing the info.  I wish you guys great success.
Somehow I suspected your project was something like this.  I will be looking
at the site in more detail as time permits.  Some time ago I wrote a
simulated trading program in Dolphin.  I called it HAL.  It used some
multi-day trend analysis to generate a collection of stocks to watch for
entry points during the day.  To spiff it up a bit I even had the little
talking wizard announce the trades as they happened.  It was fun to listen
to it calling out the orders.  I used STB to persist the data overnight.  We
started with 1 million (simulated) dollars, unfortunately it lost a few
hundred thousand dollars over a few weeks. :(   We still use Dolphin to
generate indicators, but I have put the idea of a fully automated trading
system on hold for a while. ;)  Smalltalk is great for simulation.  It would
be cool if your new site can reach a new audience of potential developers.

Chris

Andy,

> As you may have seen on c.l.s, the beans are spilled here:
>
> www.alchemetrics.org

That is a very nice site.

And a good advertisement for (Dolphin) Smalltalk -- one thing that really
shines through is a strong impression that you were having tremendous /fun/
writing those systems.

It's somewhat jarring to click through from the plushy good taste of
alchemetrics.org to the comparatively trashy-looking sites that you reference
;-)

BTW, I noticed a few small problems on the site, I mention them here on the
off-chance that you are interested....

You are missing an 'm' in 'maxi um' on the 'Money Management' page.  'Web
service' page lacks an 'of' between 'number strategies'.  On the
(entertaining!)  'No Bugs ?' page I think it should be 'experience backED up by
the TWB'.

Another small problem is that the fonts are sometimes too small for me to read
on my monitor, if I tell my browser to increase the font size then the
left-hand contents menu gets screwed-up (I'm using the latest Mozilla Firefox).

    -- chris

Bill,

There don't appear to be any problems. We don't tend to do new
development on that image; it's just for cathcing bugs (either with
walkbacks or #asserts). A couple of extra things it allows:

1) The ability to include inspectors as part of the application. These
are useful as, sort of, an under the bonnet (hood) way of looking
around the app to make sure everything is working ok. It also means
that one doesn't have to develop a GUI for important but rarely used
interfaces (remember this isn't an end-user for the masses application
so we can get away with this).

2) We don't do it yet but we will, at some point, add the ability for
the workbench to "self test" by running a subset of the test cases
before each morning's market session. I say, "a subset", because the
full suite takes a long time to run (partly because we have to launch
the broker's Java app over an over again) and partly because some tests
can occasionally fail (again due to the way we are forced to test the
broker interface).

Well I must confess that the J*va stuff has become more reliable of
late (there hasn't been a crash in the last two months) but it was a
severe situation in the early days. I'm pretty sure those crashes were
deadlocks but they demanded that the J*va machine had to be killed from
the Task Manager. And here is another issue; we have to run one copy of
the J*va app for each account being traded. Typically only one of these
would lock up but we still had to kill both (via termination of the
J*va VM) will double the resultant chaos, of course. I didn't count it
but I reckon these problems have cost us a significant amount since May.

Another reason we continue to run the J*va apps on a sole machine is
that they hog a significant amount of processing power. It is
interesting to note that our Dolphin monitoring code can scan ~2500
stocks continuously on a 3G Pentium using <30% CPU but two J*va windows
can't show 80 odd positions without eating up 80% CPU at busy times
(like the open). So much for Java being fast! We found that quite a lot
of the processing goes in repainting the price changes as they come in
so not only do we have to banish J*va to a second machine but we also
minimize the windows help the poor fellow out a bit more.
Now, I know that the Dolphin VM is unlikely to be able to take on
hotspot in a one-to-one faceoff but all this just goes to show that low
level benchmarks are pretty much useless when trying to predict the
speed of a large (or, indeed, any) application. Much more important is
the ability to design good OO code and this will likely result in good
efficiency (and refactoring will play a big part here).
Yes. IB offer a good brokerage service with an API that requires a Java
app to be launched to use it (don't ask me why). There are a couple of
other brokers with an API but they don't allow non-US residents to open
accounts.


Best regards,

Andy Bower
Dolphin Support
www.object-arts.com