PureBytes Links
Trading Reference Links
|
To All,
Just thought there might be some interest in what is going on with the
TradeLab program. The part that deals with the UMDS and interfacing to it
to store tick data may be of special interest. Especially to the
TraderWare users.
John
>Current Status
>--------------
>More than 300 user program methods have been exposed for testing, but a few
>additional methods still need to be tested before the general release.
>TradeLab documentation is nearly ready for release. There are four help
>files. They are TradeLab, TL User Programming, TL Edit, and TL Data. They
>contain more than 1,200 topics, but a few more need to be written and
>associated Lab Notes still need to be added. (This documentation is in
>addition to the extensive Visual Basic documentation provided with Visual
>Basic by Microsoft.) One additional task remains.
>
>The Primary Remaining Task
>--------------------------
>TradeLab uses MarketStream's Universal Market Data Server (UMDS) for
>real-time data. The UMDS is well designed. It is reliable. It is
>compatible with a wide range of data feeds. It is liked by nearly everyone
>who uses it. However, it was designed primarily for options traders and
>intuitive stock traders. It does not have a tick database management
>system. Because of that, it can save only a few days of uncompressed tick
>data.
>
>Long tick data histories are essential to trading strategy development and
>testing. TradeLab would be little more than a toy without a tick database
>management system capable of saving many years of compressed tick data.
>MarketStream advised long ago that UMDS 3 (M3) would have that capability,
>but it was dropped from the design. That left us with no option, but to
>design our own. That was started several weeks ago. Some portions are
>running, but it is not finished. If it had not been necessary to divert
>resources to that project, TradeLab would have been ready for release by
>now.
>
>Our tick database system is named TL Data. It is compatible with both UMDS
>2 and 3 (M3). It will be included as a component of TradeLab. It also may
>be made available to other commercial trading software developers who want
>to use it with their products, with or without TradeLab.
>
>MarketStream approves of what we are doing. They have cooperated by
>providing the interfacing details needed to make the UMDS and TL Data work
>together compatibly. We are confident that the end-result will set a new
>standard for trouble-free operation and rock-solid reliability. (A single
>bad byte can totally destroy a large database with commonly used designs.
>TL Data is designed to limit file damage due to computer crashes, power
>failures, disk errors, and other problems like that to a very small part of
>a total database in most cases.)
>
>General Release Date
>--------------------
>TL Data should be finished and testing of both it and the remaining TradeLab
>user program methods should be completed early in the second quarter. A
>more definite release date will be announced as the time approaches.
>Mailing-List: contact tradelab-help@xxxxxxxxxxxx; run by ezmlm
>X-No-Archive: yes
>list-help: <mailto:tradelab-help@xxxxxxxxxxxx>
>list-unsubscribe: <mailto:tradelab-unsubscribe@xxxxxxxxxxxx>
>list-post: <mailto:tradelab@xxxxxxxxxxxx>
>Reply-To: <tradelab@xxxxxxxxxxxx>
>Delivered-To: mailing list tradelab@xxxxxxxxxxxx
>Delivered-To: moderator for tradelab@xxxxxxxxxxxx
>From: "Doretta Smith" <doretta@xxxxxxxxxx>
>To: "TradeLab List" <tradelab@xxxxxxxxxxxx>
>Date: Sat, 19 Feb 2000 12:03:34 -0700
>X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
>Importance: Normal
>Subject: TL_ TradeLab Status
>
>Emails like these are received every day:
>
> > I heard from a beta tester that beta testing
> > is going well. When will it be available?
>
> > Can you tell me when TradeLab will be released?
>
> > I want to order TradeLab. Can you give me a status
> > report?
>
>The continuing interest is appreciated, but overwhelming. There are too
>many inquiries each day to respond individually prior to the general
>release.
>
>We have been extremely busy. Beta testing has been an intensive effort.
>Our staff has grown. Additional office space has been acquired. User
>documentation has been written. An extensive web site has been developed
>(it will not be posted until we are ready to accept orders). Order
>fulfillment preparations have been made. Many other things have been done.
>We are rapidly approaching a general release.
>
>Beta Testing
>------------
>We are fortunate to have an unusually well qualified group of enthusiastic
>beta testers. Many have taken extraordinary interest in the product. Some
>have spent an unbelievable amount of time testing, writing trouble reports,
>and making valuable suggestions and comments. Their feedback has been
>critically important to perfecting the product.
>
>In addition to their testing, we engaged the services of an outside research
>scientist to independently confirm results returned by TradeLab's extensive
>library of advanced math methods. He independently wrote C++ and Visual
>Basic shadow-code that performs the same math calculations differently for
>comparison. He also compared TradeLab math library method results to
>results obtained with Mathcad, Maple and MATLAB, where similar methods are
>supported in those programs. Anomalies were investigated in each case.
>Errors either were determined to be within the numeric precision range of
>the data types used or problems were found and code corrections were made to
>make them so. That testing has been completed for hundreds of math methods
>that will be exposed for use in the initial general release product.
>Additional math methods that haven't been tested that way yet will be
>exposed in later upgrade versions.
>
>The beta testers still have more work to do before the general release.
>Most of the testing that remains is related to trading strategy back-testing
>and optimization. TradeLab is able to much more closely emulate real
>trading during back-testing and optimization than is possible with other
>products. It also has sophisticated techniques that speed optimization,
>find better trading methods and parameters, and to determine confidence
>intervals. There is a lot of testing to do related to all that, because
>those capabilities are extensive, but we don't expect serious problems to be
>found during that testing.
>
>Even though the wait may seem endless to someone wanting to place an order,
>progress has been better than experienced developers generally would expect
>with thorough testing of a product of TradeLab's complexity. We could have
>tested like others often do and released sooner, but then the complaints
>would have been about buggy software, instead of the long wait for release.
>
>It is impossible to please everyone. If developers spend a long time
>thoroughly testing and perfecting, they are criticized for taking so long to
>release. If they merely have beta testers play around on their own for a
>while and then release, they are criticized for buggy software.
>
>We have chosen the thorough test approach, knowing that many will be
>critical of the time it takes. However, we want TradeLab to work well and
>we want to minimize technical support expense. We think those
>considerations are more important in the long run than early release, so
>that is what we are doing.
>
>Of course, no amount of testing will find all the problems in a complex new
>product. Some problems must be expected in the initial release, despite
>careful testing. End-users will use TradeLab on a wider variety of computer
>systems and in a wider variety of ways than testers have. They will think
>some things are wrong that others have thought were just fine. No amount of
>testing can totally eliminate problems like that, but careful testing can
>reduce them. That is our objective.
>
>We have used a structured beta testing method where testers have been asked
>to focus their testing on different parts of the program at different times.
>We have explained how various capabilities should be used and how they
>should work. Then we have asked the testers to test those parts, to report
>any problems, and to make suggestions for improvements. After that, code
>corrections have been made, a new build has been posted, the same
>capabilities have been re-tested, and any new problems caused by changes
>have been dealt with in most cases before moving on to something else.
>
>Structured testing like that provides several advantages. It insures that
>little used or poorly understood parts of a program are not overlooked.
>Problems can be corrected more efficiently, because rather than working on
>many different parts at the same time, attention is concentrated on
>perfecting one part at a time. Beta testers acquire more thorough
>understandings of a product, because they step through everything. Better
>understandings result in better testing. The outcome after testing is a
>pool of people who are more knowledgeable about a product and better
>qualified to answer questions and help first-time users.
>
>Current Status
>--------------
>More than 300 user program methods have been exposed for testing, but a few
>additional methods still need to be tested before the general release.
>TradeLab documentation is nearly ready for release. There are four help
>files. They are TradeLab, TL User Programming, TL Edit, and TL Data. They
>contain more than 1,200 topics, but a few more need to be written and
>associated Lab Notes still need to be added. (This documentation is in
>addition to the extensive Visual Basic documentation provided with Visual
>Basic by Microsoft.) One additional task remains.
>
>The Primary Remaining Task
>--------------------------
>TradeLab uses MarketStream's Universal Market Data Server (UMDS) for
>real-time data. The UMDS is well designed. It is reliable. It is
>compatible with a wide range of data feeds. It is liked by nearly everyone
>who uses it. However, it was designed primarily for options traders and
>intuitive stock traders. It does not have a tick database management
>system. Because of that, it can save only a few days of uncompressed tick
>data.
>
>Long tick data histories are essential to trading strategy development and
>testing. TradeLab would be little more than a toy without a tick database
>management system capable of saving many years of compressed tick data.
>MarketStream advised long ago that UMDS 3 (M3) would have that capability,
>but it was dropped from the design. That left us with no option, but to
>design our own. That was started several weeks ago. Some portions are
>running, but it is not finished. If it had not been necessary to divert
>resources to that project, TradeLab would have been ready for release by
>now.
>
>Our tick database system is named TL Data. It is compatible with both UMDS
>2 and 3 (M3). It will be included as a component of TradeLab. It also may
>be made available to other commercial trading software developers who want
>to use it with their products, with or without TradeLab.
>
>MarketStream approves of what we are doing. They have cooperated by
>providing the interfacing details needed to make the UMDS and TL Data work
>together compatibly. We are confident that the end-result will set a new
>standard for trouble-free operation and rock-solid reliability. (A single
>bad byte can totally destroy a large database with commonly used designs.
>TL Data is designed to limit file damage due to computer crashes, power
>failures, disk errors, and other problems like that to a very small part of
>a total database in most cases.)
>
>General Release Date
>--------------------
>TL Data should be finished and testing of both it and the remaining TradeLab
>user program methods should be completed early in the second quarter. A
>more definite release date will be announced as the time approaches.
>
>Additional Beta Testers
>-----------------------
>People continually ask if we are still accepting beta testers. We are not.
>More people asked to be beta testers than we could reasonably use. Many
>applications were rejected because of that. We do not plan to add more
>testers. The testers we have keep us very busy.
>
>A Pre-Release Version
>---------------------
>There have been many inquiries about the availability of a pre-release
>version. We considered that alternative and thought for a while that we
>might do it. However, doing that would divert resources and delay the
>general release, so we decided against it.
>
>Questions
>---------
>There is overwhelming interest in TradeLab. If we answered all the
>questions received each day, TradeLab might never be released. It is in
>everyone's best interest to apply our full resources to getting the product
>ready for release, rather to answering questions, so that is what we are
>doing. Questions will not be answered until we are ready to accept orders.
>
>The TradeLab web site has been completely redesigned. The new site has
>detailed product descriptions, lots of screen images, user program method
>descriptions, sample user programs, pricing, customer support information,
>information about an exclusive new LabLink mail list where advanced methods
>will be explained and discussed, beta tester comments, developer's
>commentary, and many other things. It will answer most questions. However,
>it will not be posted until we are ready to accept orders, because it would
>be counterproductive to raise the level of interest any higher than it
>already is before then.
>
>The new web site also includes a Frequently Asked Questions page that
>provides a lot of additional information. Questions can be submitted now,
>but answers will not be posted until the new web site is posted.
>
>DO NOT send questions by email. We cannot delay TradeLab's release to
>answer a never-ending stream of emailed questions. Questions must be
>submitted only via the form provided at:
>
> http://www.tradelab.net/question.html
>
>We will not be able to individually address every question, because there
>will be so many, because many will be similar, because many will not be
>specific enough, and because some would require long, complicated
>explanations. However, good questions will be selected for inclusion in the
>FAQ page. (Don't ask questions like "How does TradeLab compare to other
>trading software?" that would take a book to answer. Ask specific questions
>that can be answered in a paragraph or two.)
>
>New Company Name
>----------------
>We changed our company name several months ago. We are now known as Sanzar,
>Inc. Everything is essentially the same, except that we have grown.
>
>Who is TradeLab For?
>--------------------
>TradeLab appeals to those who want advanced capabilities not found in other
>trading software. You don't need TradeLab if you want to use the indicators
>and trading strategies used by most traders. TradeLab is designed for those
>who realize that if they are going to win, their opponents have to be
>losers. Therefore, if they use the methods used by their opponents, they
>will be using methods used by losers. Why would anyone expect to win doing
>that?
>
>No one can win consistently playing a loser's game. If you are going to win
>consistently, you have to use methods that are better than those you trade
>against. TradeLab is designed for people who understand that. It is not
>designed for people who want to know whether it has what other products
>have. It is designed for those who want to know what it has that others
>don't have.
>
>If that describes you, TradeLab will be worth the wait!
>
>-Doretta Smith, President
> Sanzar, Inc.
> doretta@xxxxxxxxxx
|