|
Paul,
I love your idea of drawing plug-ins, and I agree whole heartedly with Tomasz, as I have posted before, that I prefer a "one man band" genius programmer to a group of so-so programmers for the AB main program. Programming is a creative project like writing music -- though it is possible for others to make "arrangements" for specific instruments that add to the richness of the original tune.
When I started with AB, I used the hand drawn study tools a lot. However, I became so frustrated with the bugs, that I all but abandoned them and started writing AFL to to try and figure out how to automatically draw lines where I would have wanted them when possible.
However, there is a creative element to trading that can not be algorithmically derived within the capabilities of AFL (or my programming skills) in realtime. I know a number of traders that make very good profits trading, but they could not do it without some accurate hand drawn studies going on in realtime. That means I had to start doing most of my drawing in another program chart open side by side with my AB chart. I hate doing this and it limits what I can do with trading signals in AFL.
I almost abandoned AB at that point. However, I was attracted to the speed of AFL, and I could see that TJ was adding more capability to AFL all the time, so I stuck with it as a general purpose programming language for trading. The Gfx capabilities opened up a lot of possibilities for improving the on chart UI and I have taken advantage of this to great advantage on my charts, and adding more capabilities to the parameters window through AFL has also paid off for me.
I considered writing my own drawing routines, and I made a number of AFL suggestions that would make it easier to be able to do that, but those suggestions were not in line with TJ's vision at the time. It requires adding more support for high speed UI interactions with the user in AFL.
In reality, the studies provided in AB are almost enough to do the job. They just need to work properly. What is lacking in functionality is just a few tweaks like have already been suggested.
If a plug-in architecture were made available to completely replace the built-in studies I don't know that I could write a replacement set myself, but I am sure somebody could. I am so very close to having a full trading platform all in AB AFL at this point... just so close...
Best regards, Dennis On Jun 21, 2008, at 5:41 AM, Paul Ho wrote: Tomasz, You are pretty good to pounce on semantics. But i didnt say AB was a one man band I said Tomasz, you are a one man band. And I'm pretty convinced that I'm right by the way you replied to this post. Anyway, did you read what I say beyond that? Plugins are implemented in AB but only in AFL. I said other parts of AB will benefit greatly in that approach, in particular the hand drawing capabililties. Just to use the muic analogy a bit more because I play a bit myself. I think part of AB industralisation is more than composing music, it is conducting an orchastra to play the piece that you compose (AB) well. Hello,
For correctness sake: AB is NOT one man band.
There is Marcin doing 99% of support tasks. Plus I do outsource
bookkeeping and some other "side" tasks.
What is one-man is AmiBroker core development.
And it will stay that way for various reasons, with main reason
being that adding 2 more developers for core development
would SLOW DOWN the development and the quality
would go down hill, no matter how good these developers are.
These are facts of life. I was working in various developer teams
in the past starting from 3 people ending at 40 developers
and adding developers only adds problems. The only
way to work in a team in efficient way is to have *separate* programs
to develop (they can communicate between each other).
Sligthly less efficient (but still OK) is to have plugin approach
with well defined interface. The plugin approach is already implemented in AB.
Programming is NOT equivalent to homebuilding.
Programming is rather like composing music. At least for me.
Did you see lots of music written by more than one composer?
(even if many people appear on the "credits", usually the core idea is the work
of one man).
Best regards,
Tomasz Janeczko
amibroker.com
----- Original Message -----
From: "brian_z111" <brian_z111@xxxxxxcom>
To: <amibroker@xxxxxxxxxps.com>
Sent: Saturday, June 21, 2008 10:56 AM
Subject: [amibroker] Re: Study Charting dare I say bugs...
> Paul,
>
> Thanks for your thoughtful comments.
>
>> I also assumed that many of those who wanted/relied
>> on AB's drawing capabilities might also have deserted us already.
>
> Dennis, Joseph, Murthyuresh and Bill are still here.
> They strike me as being people with a good knowledge of AB and also
> not the types to make frivilous complaints.
>
>> Believe me, the law
>> of diminishing return applies far more rigourously in human than
>> CPUs.
>
>
> I have to agree that this is a factor but born business people thrive
> on that challenge.
>
> Small business challenges are the same all around the developed world.
>
> Example:
>
> Friends of my wife - husband and wife 6 figure+ PA turnover every
> year - earthmoving and construction.
>
> He was on mobile and got calls all day, every day and night of the
> week.
> He was a one man 'operating' band - she 'did the books' and was the
> hidden business brains.
>
> I told him "if you are available 24/7 they will call you 24/7, if you
> are available 9 to 5 they will call you 9 to 5 and you will still do
> exactly the same amount of business".
>
> Also - the decision to change from a 'one man band' operation to
> a 'business model' is very hard for anyone and excruciating for some.
>
> I told him "you have to learn to trust people and delegate and then
> you can get some fishing time".
>
> Today his mobile is private and only for family use.
> He has a business (sort of) manager and an office (sort of) manager.
>
> I told him "Now you have to give your 'manager' his own project to
> run from start to finish and give him a % of the profit.
> If he is successful then give him the next project as well and take
> your first holiday in 10 years" (his wife always listens quietly).
>
> Then I said "It doesn't matter what you do the staff will never do it
> exactly as you will do it yourself and they will never be quite as
> good at it as you are".
>
> He turned to me and said "Yes, but they are good enough".
>
> His business has doubled in the last 3 years.
>
> His wife is a very good student.
>
> brian_z
>
> --- In amibroker@xxxxxxxxxps.com, "Paul Ho" <paultsho@xx.> wrote:
>>
>> Being a patron of AB for more than 4 years. I would say that AB is
>> very much an AFL centric product. The changes in the AFL engine and
>> its capabilities has been nothing short of phenomenal. In contrast,
>> the hand drawing features of AB has hardly changed in the last 4
>> years, and probably not much has changed for much longer than that.
>> Although not publicly stated, I presumed that it is the direction
>> that Tomasz wants to go, and this specialization has been very
>> successful. I can also assumed that many of those who wanted/relied
>> on AB's drawing capabilities might also have deserted us already. I
>> have known some of the problems Dennis has mentioned long time ago,
>> but somehow I have never bothered even to mentioned them, probably
>> because I developed myself more on AFL as a trading and charting
>> tool, and hardly use the drawing tools these day.
>>
>> Tomasz is a one man band. So I can apprecriate and am very happy
> with
>> the product direction he has taken. I dont think Brian's suggestion
>> of hiring a few more programmers is the answer. Believe me, the law
>> of diminishing return applies far more rigourously in human than
>> CPUs. What I would like to see is that Tomasz opens up the hand
>> drawing capability to third party contribution throuh the use of
>> plugins, much like the AFL and data plugins. Once done, I'm sure
> that
>> people will roll their own drawing tools. Of course, existing bugs
>> will still need to be fixed.
>>
>> --- In amibroker@xxxxxxxxxps.com, "brian_z111" <brian_z111@> wrote:
>> >
>> > > And no, I can not hire 2 more programmers because then I would
>> need
>> > >to rise the cost of the program 3 times
>> > > and then everyone would scream that it is too expensive and
> will
>> > >walk away. Then it would mean the end of product.
>> >
>> > Since Dennis's intent is to invite constructive comment and since
>> the
>> > forum does act as an informal customer feedback center:
>> >
>> > the content of your reply relates to a lot of personal and
> business
>> > decisions - not sure if you are inviting comment on those matters?
>> >
>> >
>> >
>> > As a customer:
>> >
>> > I am not certain that I am a price buyer (once again this relates
>> to
>> > business theory and application), probably not.
>> >
>> > The price is nice but is it the reason I buy AB?
>> >
>> > In the first place I made a choice to make trading my home based
>> > business (I could have chosen something else), so that makes me a
>> > business customer I guess (albeit a small-time one).
>> >
>> > You can't trade professionally with anything less than approx
> $100K.
>> > If I invested that amount with an Aussie fund I would pay around
> 2%
>> > expenses, so 2K PA is a reasonable operating cost.
>> >
>> > When I first started my costs were high (books, training courses
>> etc)
>> > but now I can operate around those figures (I do my own tax and
>> only
>> > need software/data/computer/internet - but I would have
>> > computer/internet anyway - I don't need specialist computers for
>> > trading).
>> >
>> > Of course that is only if we don't consider brokerage as a
> business
>> > cost - if I daytrade 100k what is my annual brokerage bill? -
>> > mentally I lump this in with gross edge - commissions == net edge
>> but
>> > in business terms it is on the books as operating costs.
>> > Software/data costs are a tiny fraction of annual commissions.
>> >
>> > Anyone who has operated a business, in one of the major economies
>> of
>> > the world, will know that trading costs are laughable compared to
>> the
>> > costs of running other businesses - most businesses in Australia
>> run
>> > on very low net returns compared to the capital invested in the
>> > business.
>> >
>> >
>> > Also when I bought AB I was not a naive trader.
>> > I knew what I was doing and had owned two other software packages
>> > prior to AB.
>> >
>> > On that basis I would say I am a specialist customer (perhaps
> that
>> is
>> > AB's niche in the market).
>> >
>> > Specialists want the software to do the things they want it to do
>> > (pretty hard since we all want different things).
>> >
>> > Adding all of that up I vote for:
>> >
>> > please hire another programmer, give people more of what they
> want
>> > (give the optimizers their dual core stuff, give the chartists
>> their
>> > snap and draw or whatever, give the Aussies their YahooFdata etc,
>> > etc) and put the price up - maybe even add some features
>> > to the UKB :-)
>> >
>> > Metastock is the point of comparison for AB - it is approx $500
> PA
>> > for EOD and from $1000 for Pro - so AB can stand that price (of
>> > course they have 'lease' MS for a monthly fee which allows us to
>> > pretend we are not paying $1000 a year and so that makes it feel
>> > nicer).
>> >
>> >
>> >
>> > Re price buying:
>> >
>> > If I buy a cheap car I get less features.
>> > If I pay more I get more features and better appointments.
>> > If I buy the cheap car I still expect the (lesser) features to
> work.
>> > If they don't I won't buy it again at any price.
>> >
>> > (if an existing feature doesn't work this has a far more negative
>> > effect on customers than not having the feature they want in the
>> > first place? - psychologists claim that one negative experience
> far
>> > outweigns many positive experiences - also it is human nature to
>> buy
>> > the cheaper model and then complain because it doesn't have the
>> > features they didn't pay for - that will never change).
>> >
>> >
>> > Re: walking away
>> >
>> > When it comes time for you to buy a new car what do you do Tomasz?
>> >
>> > Are you going to stick to a brand, out of loyalty, if a competing
>> > model has better reliability, features and appointments at the
> same
>> > price?
>> >
>> > How about availability, delivery and service?
>> >
>> > What motivates you to buy a product and then rebuy it over and
> over?
>> >
>> > Also you will find your needs change with time - as the kids grow
>> the
>> > house and the cars get bigger and after they go you and your wife
>> can
>> > have an office each (they used to be the kids rooms).
>> >
>> > So, perhaps people will walk away if their needs change and/or AB
>> is
>> > no loner meeting their needs.
>> >
>> > - better ask your wife, about the car, before you buy though -
> it
>> > always pays to get a second opinion ;-)
>> >
>> >
>> >
>> > Re: what do customers want?
>> >
>> > We want it all - don't you?
>> >
>> > That will never change either.
>> >
>> > The more that companies meet those 'impossible' needs the more
>> > customers they get.
>> >
>> > For example - who goes to Starbucks to pick up a coffee, then
> goes
>> to
>> > Wendys to pick up a donut and takes them to the burger shop to
> put
>> > them together for a nice breakfast if they can get a nice burger,
>> > coffee and donut at Maccas (plus a free paper to read as well)?
>> >
>> > - better get a second opinion from the kids on that one too ;-)
>> >
>> >
>> >
>> > Hope my perspective offered some helpful feedback.
>> >
>> > Regards,
>> >
>> > brian_z
>> >
>> >
>> >
>> > --- In amibroker@xxxxxxxxxps.com, "Tomasz Janeczko" <groups@>
>> > wrote:
>> > >
>> > > Hello,
>> > >
>> > > You people don't seem to realise that I am THE ONLY PERSON who
>> > works on AmiBroker code.
>> > > The day has only 24 hours and one person has only two hands.
>> > > Now if even few people ouf of 8000 users submit "ideas"
> everyday,
>> > no matter what I do I can not implement
>> > > them IMMEDIATELLY.
>> > >
>> > > And no, I can not hire 2 more programmers because then I would
>> need
>> > to rise the cost of the program 3 times
>> > > and then everyone would scream that it is too expensive and
> will
>> > walk away. Then it would mean the end of product.
>> > >
>> > > So, considering that I am the ONLY DEVELOPER of AmiBroker and
>> wrote
>> > everything from scratch with my own hands, I can say that:
>> > > a) I am doing more than anyone in this business to make
> customers
>> > happy
>> > > b) I am implementing MORE features/user requests/ideas than
>> anyone
>> > else in this business
>> > > c) I am going an extra mile with responding PERSONALLY on the
>> > public mailing list to many of your posts.
>> > >
>> > > I can not do MORE than I am doing now. It is physically
>> impossible
>> > for anyone.
>> > >
>> > > So, if I tell you that I am going to check things, I will do
>> that,
>> > but this is all subject to TIME and PHYSICAL ABILITY constraints.
>> > >
>> > > Pushing the string does not help at all. It makes things worse.
>> > >
>> > > Best regards,
>> > > Tomasz Janeczko
>> > > amibroker.com
>> > > ----- Original Message -----
>> > > From: wavemechanic
>> > > To: AmiBroker, User
>> > > Sent: Friday, June 20, 2008 2:47 PM
>> > > Subject: Re: [amibroker] Study Charting dare I say bugs...
>> > >
>> > >
>> > > Dennis:
>> > >
>> > > I completely agree with your overall assessment and know that
>> at
>> > least several users also completely or partially agree but may or
>> may
>> > not respond - so you are not alone. Much has been said through
> the
>> > years on the Yahoo board/AB website/support mail about
>> > bugs, "clunkiness", and desired/necessary features. Because of
> the
>> > type of problems that you mentioned (I am also indicator mode
>> > oriented), I use the EOD version of AB and then only at "half
>> > throttle" because my applications and programming needs are not
> as
>> > extensive as that of a number of other AB users. Like you, I use
>> > broker supplied and personal software for RT streaming and
> charting
>> > studies (no "shoe" fits all). These issues, however, are not
>> > restricted to AB as evidenced by a recent review on another board
>> > where the user compared AB/Metastock/Tradestation/Blocks and
> ended
>> up
>> > (as I understand) with a Blocks package that includes
>> > charting/streaming data/EOD data/backtesting/useful standard
>> > features - so things are evolving.
>> > >
>> > > Bill
>> > > ----- Original Message -----
>> > > From: Dennis Brown
>> > > To: amibroker@xxxxxxxxxps.com
>> > > Sent: Friday, June 20, 2008 12:12 AM
>> > > Subject: [amibroker] Study Charting dare I say bugs...
>> > >
>> > >
>> > > Friends,
>> > >
>> > > AmiBroker is a wonderful program for many things. In fact
> I
>> > use only
>> > > a fraction of its total capability for my trading style. I
>> > trade only
>> > > single ticker (futures) real time and write whole trading
>> > systems just
>> > > in AFL in indicator mode. This is actually more like the
>> > traditional
>> > > chart trading that is offered by many brokers.
>> > >
>> > > I would expect that AB, which supports all these functions
> as
>> a
>> > very
>> > > basic operating mode would be at least as good as other
>> > charting
>> > > programs.
>> > >
>> > > However, even though the fundamental operations are there
>> (and
>> > I think
>> > > potentially better), the interface (either design or
>> > implementation
>> > > bugs --I don't know which) has made it hard for me to use
>> this
>> > product
>> > > for drawing even basic lines on the charts while trading.
> In
>> > fact, I
>> > > need a second chart from my broker's application open to
> the
>> > same
>> > > ticker just so I can draw lines where I want them even
> though
>> > AB would
>> > > be superior at this if it were not for the bugs. This has
>> been
>> > going
>> > > on with me for over a year now and it is costing me.
>> > >
>> > > I have other trading friends that I have recommended try AB
>> in
>> > the
>> > > last week, and I am embarrassed to think what they will say
>> to
>> > me when
>> > > they figure out that these basic features don't work
>> properly.
>> > I have
>> > > actually refrained from recommending AB to them for quite a
>> > while --
>> > > hoping these things would be fixed first.
>> > >
>> > > I don't know how many other traders are frustrated by the
>> same
>> > bugs,
>> > > but if you are, please join the conversation so that Tomasz
>> > knows that
>> > > this is important to you also and perhaps we can get some
>> > higher
>> > > priority on getting the study charting to work properly.
>> > >
>> > > Here is a short list of my frustrating bugs --there are
>> others
>> > on the
>> > > suggestions and bug lists, but I am only pointing out the
>> ones
>> > that
>> > > cost profits by either making it very slow to do a simple
>> > operation
>> > > that makes one lose the trade opportunity, or creates
>> > frustration and
>> > > emotion which takes one out of the trading "Zone", or makes
>> the
>> > line
>> > > drawn correctly to become inaccurate causing one to miss
> the
>> > trigger
>> > > point for a trade. If I have missed any that you feel are
>> > costing you
>> > > profits (not the nice to have pretty things) than please
> add
>> > them to
>> > > the conversation.
>> > >
>> > > #726. Handles from a selected study line are not
> respected.
>> > Often
>> > > several lines are anchored or pass through the same pivot
>> > points.
>> > > Price levels, Trend lines, Forks, Fibs, etc., often use a
>> > common pivot
>> > > point. However, as a trade progresses, new pivots become
>> more
>> > > important and the lines need to be adjusted in real time.
> It
>> > is easy
>> > > to select a study just by clicking on it in a spot where it
>> is
>> > unique,
>> > > and the handles will appear along the line. However,
> unlike
>> > any other
>> > > charting program I have ever used, clicking and dragging on
>> > the
>> > > visible handle will not usually work properly if any other
>> > study
>> > > crosses very close to that handle. The problem is that AB
>> will
>> > grab
>> > > the earliest drawn study and not the selected one --it will
>> un-
>> > select
>> > > the one you selected and select the oldest one and drag it
>> > instead.
>> > > After playing around for a few valuable seconds trying to
>> grab
>> > the
>> > > right line, I end up having to move the whole line away
> from
>> > its
>> > > current location and reposition all its handles. Very
>> > frustrating and
>> > > takes the time and concentration away from the trading
> action.
>> > >
>> > > #728. Horizontal price levels do not stay horizontal. The
>> > horizontal
>> > > lines (price levels) do not snap to a bar with the magnet
>> mode
>> > on.
>> > > They just are placed at the cursor level. However, if you
>> > click on
>> > > the line, it jumps to an angled line that snaps to the bar
> on
>> > the left
>> > > side of the screen. Its operation is just bizarre and
> causes
>> > errors
>> > > in placement of the price level after the fact.
> Completely
>> > > unacceptable behavior for trading.
>> > >
>> > > #890. Study handles change when off screen. Lets say you
>> have
>> > the
>> > > magnet on and are zoomed out and draw a trend line between
>> two
>> > peaks.
>> > > Next you zoom in on the trading action so that only the
> right
>> > hand
>> > > handle is visible on the chart. Now say the prices have
>> > advanced to a
>> > > new peak and you need to adjust the right handle to the new
>> > peak (this
>> > > is a very common occurrence). Most of the time, that
> handle
>> > that is
>> > > off screen jumps to a new bar position and is no longer on
>> the
>> > peak
>> > > you selected. Of course that changes the slope of the
> trend
>> > line to
>> > > the wrong one, but you never know it when trading and you
> end
>> > up
>> > > placing bad breakout trades because the line is in the
> wrong
>> > place
>> > > now. Completely unacceptable behavior for trading.
>> > >
>> > > #990. Drag shadow is on the wrong bar. When dragging a
>> study,
>> > there
>> > > is a thin "shadow" line that shows which bar you have
> dragged
>> > the
>> > > handles to. However, when the mouse is released, the study
>> > gets moved
>> > > to a different bar than the one indicated by the shadow.
>> This
>> > wastes
>> > > valuable time with two or three attempts to figure out
> which
>> > wrong bar
>> > > to put the shadow on to get the study to go where you want
>> it.
>> > This
>> > > is an annoyance.
>> > >
>> > > And here is one more that I just found today. If I draw
>> trend
>> > lines
>> > > on a one minute bar snapped to peaks, then switch to a 5
>> minute
>> > view,
>> > > the trend lines are shown on the wrong 5 minute bars. And
>> > heaven help
>> > > me if I select the line in that view, because it will snap
> to
>> > these
>> > > wrong bars.
>> > >
>> > > There you have my short list. Please speak up if these
>> issues
>> > affect
>> > > you.
>> > >
>> > > Best regards,
>> > > Dennis
>> > >
>> > >
>> > >
>> > > ------------------------------------
>> > >
>> > > Please note that this group is for discussion between users
>> > only.
>> > >
>> > > To get support from AmiBroker please send an e-mail
> directly
>> to
>> > > SUPPORT {at} amibroker.com
>> > >
>> > > For NEW RELEASE ANNOUNCEMENTS and other news always check
>> > DEVLOG:
>> > > http://www.amibroker.com/devlog/
>> > >
>> > > For other support material please check also:
>> > > http://www.amibroker.com/support.html
>> > > Yahoo! Groups Links
>> > >
>> > >
>> > >
>> > >
>> > >
>> > > ----------------------------------------------------------
> --
>> --
>> > --------
>> > >
>> > >
>> > >
>> > > No virus found in this incoming message.
>> > > Checked by AVG.
>> > > Version: 8.0.100 / Virus Database: 270.4.1/1510 - Release
>> Date:
>> > 6/19/2008 3:21 PM
>> > >
>> >
>>
>
>
>
> ------------------------------------
>
> Please note that this group is for discussion between users only.
>
> To get support from AmiBroker please send an e-mail directly to
> SUPPORT {at} amibroker.com
>
> For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
> http://www.amibroker.com/devlog/
>
> For other support material please check also:
> http://www.amibroker.com/support.html
> Yahoo! Groups Links
>
>
>
Please note that this group is for discussion between users only.
To get support from AmiBroker please send an e-mail directly to
SUPPORT {at} amibroker.com
For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
http://www.amibroker.com/devlog/
For other support material please check also:
http://www.amibroker.com/support.html
![]()
__,_._,___
|