Join Bridge Winners
All comments by Nicolas Hammond
You are ignoring the author of this comment. Click to temporarily show the comment.
Leo,

It's not ready for testing, in the sense of how everyone understands testing. If ACBL ever does anything with it, I'll pass on your name as a tester.

It's ready for a review by independent third parties. ACBL never gave it a fair review.

Even now, the people I have added as reviewers will not have been given a demo so they are doing this somewhat blind.
March 15, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
For the reviewers, I have opened it for anything.

My claim is that within 4-5 days it was ready for release to 2-3 TDs to start running KOs. This is the code that ran Gatlinburg 2014. Certainly not perfect, but you have to remember the state of the contract/project back then.

What would be great if they gave their honest opinions of how this compares to the current ACBLscore.

ACBL initially claimed it was all throw away. And wrote off all this money.

My claim is that the software worked, it had been in a state of needing feedback from users for about a year.

I haven't looked at this software for about 10 months. After the contract was over, I worked on improving some of the code/adding new features, but as this wasn't under contract, I haven't given anything to ACBL. This code is now Bridgescore+.
March 15, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
It is known that ACBLscore does not implement the rules correctly. I won't get into specifics.

I have automated test tools that can take an ACBLscore game file, import to ACBLscore+, re-score, re-rank, re-masterpoint, compare ACBLscore+ MPs with ACBLscore MPs.

The error rate for the tournament I primarily used for testing was about 10%, i.e. about 10% of game files had incorrect MPs based on the published specs. Most of this was user error, i.e. TDs incorrectly configuring the game, some was ACBLscore not matching the published specs (“well, it was easier to code it this way rather than the specs they wrote”).

We did lots of testing with lots of game types, e.g. club data for a year, ACBL gave me access to lots of tournament data.

Going forward, I think ACBLscore+ should match the specs. However something as simple as working out the averaged number of masterpoints on a team (e.g. add all player MPs, then divide by the number of players) needs to be defined. In some cases, ACBLscore+ gets a different answer than ACBLscore, which will then lead to different MP values for the event. Yes… it gets down to that level - ACBLscore, based on Pascal math libraries, and implementing its own version of rounding, combined with the Pascal math library rounding (banker's rounding) can generate different results than ACBLscore+ even with the same written specs!!

I think if we explain away the 0.01 errors as rounding (and it can get as high as 0.05 because ACBLscore rounds, then multiplies, then rounds again), that players will be happy. Seriously, when was the last time anyone verified the MPs they won?

I spent way too much time on this problem. My view is that we should come close, but not need to match. The rounding issues were a huge time-waster.
March 14, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
I've posted a separate thread for ACBLscore+ reviewers.

I've sent email with URL/passwords to Ed, Adam, Kevin, Larry, Greg, Jay.

March 14, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
Ed: OK.

My internet connection where I am is bad; it failed after a partial upload. I've restarted it. I am out all today, probably not going to get back until around 10pm EST.

I'll finish the upload, sent you all a private BW email, and will start a new thread.

Guess I'm forming my own CEO Technology Committee…

You guys can all give me good insight into what you want. We can make it public so can share with ACBL. You can also give an idea of how far along you think the software is, warts and all.

I have nothing to hide with any of this stuff.
March 13, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
I haven't seen the video. I haven't heard the audio. The first two “project managers” assigned to be my ACBL co-ordinators are no longer with ACBL.

I would consider this a small/medium project; not a large one. It may be for ACBL. It depends on your background/experience. It's only 200-250K lines of code.

I smile when I see “get feedback from stakeholders”, FD etc. They were all in ACBLscore+.

It took 5 months to get the contract negotiated, a large part of that was FD.
March 13, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
OK.

I'm traveling. My Internet connection is much, much slower than Greg. I haven't watched the video, probably won't have time/Internet access.

I'm uploading the ACBLscore+ code that I have. It's slow ;-(. I have various other things on today, including meetings late afternoon/early evening. I've started the upload to my Internet site, but it probably won't be until tonight that I get it done and set up.

I'll add Kevin, Larry, Adam (no last names necessary??). Robb, you too. I'll send you private BW email with URL/password.

If there is anyone that is going to be at the BOG meeting, that has the appropriate background, I can add you too.

Greg: if anyone on your committee wants access, let me know.

Jay: I'll send you email about this site; to the best of my knowledge no-one on your committee has seen this either.

I'll put up a page where you make comments; I don't want to hijack this thread.

Remember: this was intended as a release to TDs. It's not perfect. Never claimed it was, but take a look at where the code is/was, so you can see how close it is.
March 13, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
Anyone interested in seeing ACBLscore+?

I can probably put a version on-line. I would limit the number of people with access so the server doesn't get bogged down.

You can see what all the fuss is about.

You may be able to help the Technology Committee.
March 13, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
The contract ended in March 31, 2014. We gave ACBL a 30 day notice (with 30 day grace) in January 2014 of material breaches of the contract. Those breaches weren't fixed. So we mutually agreed upon March 31, 2014 as an end date. There was money left in the contract. We had done everything, including parts of the final phase except for any item that had a documented 3 month delay caused by ACBL. We sent a final status report (it's stated as such in the status report), along with the code.

We started negotiating a new contract.

ACBL saw what happened in Gatlinburg.

They wanted this code. They wanted to change the end date. We agreed. We sent them the code, they sent us a check.

At this point, with all money spent, ACBL decided to “terminate the contract”. What that meant is that they released me from an obligation to be available to them for so many hours per month for about a year. If they didn't want me that month, they didn't pay me. So with the contract ended there was no benefit, other than a marketing one, to “terminating the contract”. So Robert can stand up there and say “we terminated the contract”. No. HS terminated the contract first, then the contract ended. There was nothing for them to “terminate”, except release my from my time obligation.

Why?

The sticking point in the negotiation was ACBL's insistence on re-writing the original contract and making it a work-for-hire so they would own the copyright.

Why?

ACBL owns the code. HS has a full license to the code. For code developed before the contract started, HS owns the code, ACBL has the full license. Full license= no royalties, do what you want etc. But in all cases, HS owns the copyright.

I have never disputed or claimed anything to the contrary.

For some reason, when ACBL hired outside counsel to review the technology contracts, the copyright was a sticking point. I believed they hired OC for LTPB contract. See that site, ACBL has the copyright. The OC reviewed existing contracts, including ACBLscore+. This was in 2013. It's why ACBL stopped paying invoices; they wanted to change the contract in mid/late 2013 and simply stopped paying.

ACBL outside counsel stated that without the copyright, ACBL could not protect themselves.

So… if you ask the question

“Does ACBL own the copyright to the ACBLscore+ code?”

You should get a yes/no answer.

Instead you will get the answer

“ACBL owns the code.”

So…. you think that you have been given the answer to your question. You haven't.

Try it at the BOG meeting. It's a yes/no question.

“Does ACBL own the copyright to the ACBLscore+ code - yes/no?”

If you get the answer, “ACBL owns the code”, try again until you get either a Yes or a No.

Then you ask, “Did ACBL ever get any advice from outside counsel on using ACBLscore+ based on copyright?” That should also be a yes/no answer. You'll probably get some answer about legal privilege instead.

Getting a straight answer to those two questions will explain a lot of things. You've been given the run-around for too long.

March 12, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
The IKEA analogy is a good one.

For Greg, I took the code from the DVD I sent to ACBL on May 28, 2014. I typed:

% bash setup.sh
% rails s

The setup.sh script installs the code, the rails s starts it. It works.

That would have worked back in May 2014 on a current system.

Since then there has newer software release, so you have to make small changes (I sent them to Greg) so that the setup.sh script works to install older software.

The entire process takes about 3 minutes; most of the time is loading the member database into the system.

This installation process has been the same probably at least a year before the contract ended.

If ACBL were downloading regular copies (as they were supposed to), this would not have been a problem.

I took the same code shipped to ACBL last year, copied it to my Internet machine. Same two lines (same change), installed on a Linux machine. Took 3 minutes. No extra work needed. There's one line you type to make it a production system. It takes 3 minutes to run.

In the space of 10 minutes, I had ACBLscore+ running on my Mac laptop, and two instances on the Internet. All from the DVD sent to ACBL a year ago.
March 12, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
In the underlying Pascal, there are checks for club_mode or tournament_mode.

We implemented the same concept in ACBLscore+. The main reason is familiarity. To get acceptance, it's a new UI, but the closer we can keep it to the old one the better. The other main reason is support. Club support and Tournament support are very different. You want to minimize the number of unfamiliar menu items you show a club user.

Absolutely one application (in my opinion). Same as ACBLscore. Just one download.

The internals for ACBLscore+ tournament mode and club mode are the same. The difference is the parent of the event (in the implementation in the DB). It's either a Club or a Tournament.

DBADD has gone. There is no need for it in the new software.
March 12, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
“I think it's great you guys are working things out.”

I was very careful not to say that. My relationship was ACBL was a business relationship. It ended. I bear no will. It's business. It happens. Still doesn't mean I don't want the project to succeed.

“.. 41st floor ..”

That's the Technology Committee meting, not the showdown. I still plan to do it on Wednesday. I just need to find a location. Will post here for those interested.

“technology testbed”

Don't remember using those words,

“technology fanatics and the user was going to be sold down the river”

Anything but. I'm a pragmatist. I figured that the software needed about a year of alpha/beta testing/feedback before it's ready for a final release, i.e. involve the user. I still believe that. Doesn't mean large parts can't be delivered before then, but I now have enough practical experience that delivering the cheapest, best quality product is around that time frame. Could get it done in ~ 6-7 months, but would be more expensive and not as good. Considering the pick 2 of 3: time/money/quality, I would pick money (cheapest) and quality (best). Let time slip, but make sure that there are deadlines so this is tracked.

I talked to the developer today who worked on movements until he was furloughed - he estimated around the same amount of (elapsed) time - he's retired/part-time. To get this finished, movements is critical path. So is masterpoints. Got a lot of masterpoint code done, have been since ~ 1 year into contract. What is still missing is the assignment/eligibility rules. I never got them. Should be about 2-3 pages. I've reverse engineered some. I have yet to meet a TD that knows more than 50% of these rules.

My original view on masterpoints was that ACBLscore+ needed to match ACBLscore 100%. Even down to the 0.01 level. It was a noble and correct goal. We spent a lot of time on it.

Now I'm at the point where I don't care about 0.01 discrepancies. For larger MP events, I don't care about 0.05. Sounds bad. But ACBLscore doesn't implement the specs in all case, plus I can't even get something as simple as add 4 players MPs together, divide by 4, and get the same results as ACBLscore! (There are edge cases where we disagree because of the mathematical logid in ACBLscore).

For the missing eligibility/assignment rules, I may follow someone's suggestion of just implement what I think is right and let someone challenge it later. That approach has worked “for centuries”.

So I've gone from the “right” approach to the practical approach.

“then all of a sudden next month was a total wreck”

Perhaps I should release a copy of the code delivered to ACBL, put it on the Internet (I have a license for it), and solicit feedback from users. This way you can see where it's at. I never claimed it was finished, I do claim that within a week we can put something up where ACBL can start running KOs using the software (in parallel, to get training/testing/confidence) then drop ACBLscore. That's the version that ACBL has. The version that I have (Bridgescore+) I could do tomorrow; in fact I did do a release back in September to interested districts. The version I would put up could be reviewed by those with some knowledge of the software it's replacing (sorry Greg/Uday), so that you can see where the future could have been.

I've stuck to the same story since I first wrote something. No-one has yet proved anything I've written as false.

March 12, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
“Let's try again. Can you run a full KO from start to finish, over the course of two days? Including masterpoint calculation and reporting to the main office? That's what we asked you on the phone. If you have a different answer now than you did a couple of weeks ago, say it here. I wrote this down in my notes and circled it, so I don't really see how we could have miscommunicated, but anything's possible, I suppose.”

Yes it can. (ACBLscore run KO start to finish with masterpoint calculations and report to main office).

My answer is still the same - yes. There was miscommunication. Greg/Uday were confused on how masterpoints are reported/recorded. It's difficult for them because they don't know ACBLscore works, nor how Horn Lake internally processes the data.

Greg/I covered this in a subsequent phone call. He's posted something since.

As part of this, I looked over the ACBLscore+ code as delivered to ACBL back in May 2014.

Where the ACBLscore+ stands, it probably needs about a week of clean-up, if that, including removing all pair/Swiss menu items then it is ready to be released to TDs for running KOs. This would start the feedback process for TDs. The iterative feedback process is the most time consuming element towards a full final release. ACBLscore+ is behind where it should be for the feedback process. But within a week, I'd release it for KOs only. Start to finish.

It took me 3 minutes to install ACBLscore+ on my Mac, Greg had problems because he was running newer versions of many software. I took the same original ACBLscore+ source code and installed on the internet - same 3 minutes. (Takes 3 minutes to import the member database and create some test data suitable for a demo). It took 3 minutes to go from the version on the Internet to create a production version. So, in 10 minutes, I had 3 different installations - one on Mac, one on Internet (for testing/development), one on Internet (for production).

I would recommend releasing a version that supported KOs only (I have done that with Bridgescore+), get TDs to use it, initially run in parallel with ACBLscore, very quickly ACBLscore would no longer be needed to run in parallel as TD were confident in its use. ACBLscore is always available as a backup. Get feedback from TDs, incorporate into the code. Iterate. Release more features over time (e.g. Swiss).

I just wanted to clarify here, as Greg had posted a specific question that appeared unanswered.
March 12, 2015
Nicolas Hammond edited this comment March 12, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
That's a step up from when I saw you getting the information from the Jeffrey's chart!

Unfortunately the DIC computer rarely has Internet access so having him/her upload th e data won't work either.

You could find another volunteer to get the information for you; but then they have to find you so that doesn't help much.

I could set up a site so someone can upload the file and the results would be on an Internet site i have and you could copy/paste from that, but we'd have to find the person; it's not much extra work for me, but seems a bit of a hassle for all involved.

The current process, bad though it is, is probably the most practical.
March 12, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
A big personal public thank you to Robert Hartman and Greg Humphreys for their help in the last 2-3 days.

There is often a lot more happening in private than is discussed in a public forum.

Many of you have made unfounded, public speculation that really isn't helpful to the discourse, but we ignore you.

This doesn't mean I retract anything I've written; or am changing my position, but I want to publicly acknowledge that both Robert & Greg have gone beyond the call of duty the last couple of days.
March 12, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
Jan,

We've chatted about this before.

I think I even gave you the results from the Atlanta NABC Swiss in 2013 during the event, because was quicker/easier for me (I was running software in parallel).

If you can get a copy of the Game File from the director (bring a USB stick, they are used to backing up game files to a stick), it should be easy. It's a quick operation for them. It doesn't need a printer.

I've got some (free, available to all) tools on the Internet to extract data from a game file. It will only be a little bit more tweaking for me to take the game file and then print out the current leader board. You upload the file, it sends back the leader board. Then you will only need to cut/paste from one browser window to another browser window. It will be a semi-automated process. It will save a lot of the time that I currently see you typing in the player names.

If you get permission from the NABC+ Swiss DIC, let's chat during the week (I get there on Monday) and we will do a couple of dry runs. I think you have my cell #.
March 11, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
Because that's the way it has always been.

I suspect Jim added a club/tournament mode to distinguish between which MP options are displayed, i.e. you have to a tournament to access menus to create a Regional event.

I suspect that colors were added to help support calls. Rather than “are you in club mode or tournament mode”, the question is “is the background red or green”. The latter is much easier to answer.

Same thing with ACBLscore+. In creating an event for a club, you can't create a regionally rated event.

My support plans for ACBLscore+ were a little different than the current ACBLscore; but that's history.
March 11, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
The club/tournament mode in ACBLscore just affects the color of the background screen and the list of menu items displayed.

We have the same concept in ACBLscore+. If working with clubs we show a different menu. I didn't bother with a different background color.

But the underlying code is the same. The UI for ACBLscore+/Bridgescore+ is the same - we don't change background color.

in Bridgescore+/ACBLscore+ there's more features for the DIC as well.
March 11, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
“But I AM accusing Nic of NOT breaking his work into smaller projects. I don't care who your customer is. It's mandatory”

It was.

“Maybe Nic asked for a site in which he could place his latest beta version of ACBLScore+ for downloading and comments. I would have”

I didn't ask. I just did it.

I created a site for outside comments on ACBLscore+. May 2012. One month into the contract. Exact purpose was for club owners/managers/players to submit suggestions.

Separately there was a site for ACBL TDs to provide comments (didn't want to make this public).

Also, I create a site to demo ACBLscore+ at any point to ACBL, the BOD etc. etc. It was up to them who they gave access to.

“A beta version should have been available, and mentioned on the ACBL home page. A beta program is important, so that club managers like myself can see what Nic is up to and comment as he goes along.”

My plans were to start with an alpha version. 3-10 clubs. I had the clubs lined up. I wanted to do an alpha release one year into the contract after a successful demo in Gatlinburg 2013. I figured that a year of comments/feedback from clubs was about right for this type of project. ACBL said no. I still think that about a year of comments/feedback is about right.

“Maybe Nic asked to post his schedule and specifications”

I did ACBL to do certain things to help with education.

At the same time, there was a project web site that was open for comments.

“Nic says, “I'm sorry, but clubs are at the end of the list.” Clubs are all I care about.”

Whoa. Let me be very clear:

When I was working on ACBLscore+, clubs were a huge priority. I wanted an alpha release to clubs a year into the contract, and this was _before_ the TDs would have a version. There's more to learn from the clubs than the TDs. At the time this made sense, we had developers that were club owners/directors and so could ‘swim in their own muck’ so to speak. We also had other clubs that were a short driving distance from the developers. Support is crucial for an alpha release. You keep it private as there will be lots of bugs/issues to work out.

However, the contract is now over. I don't work for ACBL. I'm working on Bridgescore+.

If I were to take over ACBLscore+, in its current state, I would recommend a release to TDs, then a release to clubs. No offense, but the tournament side is currently further along than the club site of the code, mainly because the tournament version has had more exposure and testing than the club version.

“secret project manager”.

It was not a secret. ACBL even gave me a badge to wear at tournaments.

“There was no deadline for the first version. It was maybe 1 to 2 years away”

There was.

I think you keep forgetting that in January 2014 I had to put ACBL on 30 day notice to correct material breaches in the contract. After a 60 day period when these were not corrected, my company terminated the contract.

Larry: I'm actually on your side. There are some things on your wish list that were done during the ACBLscore+ contract. I see no reason for ACBL not to make these available to club owners/managers/directors. I'm telling you what they are, how to get them.

I'm trying to get ACBLscore+ to have a fair review. I think ACBL should proceed with it. That's my opinion. They can do it with me, or without me. I believe in the ACBLscore+ code - the design, the implementation etc.. I've taken what was ACBLscore+, worked on it without the restriction on ACBL for the last year, and added many features and called it Bridgescore+. Everyone who has seen it or used it seems to like it. I'm asking for ACBLscore+ to have the same chance.
March 11, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
Kevin:

Gatlinburg is meeting highly unlikely. Pete has been great to work with, lots of his suggestions are incorporated into Bridgescore+. Same with Doug Grove, the now retired DIC of Gatlinburg. Charlie McCracken, the new DIC, has also made some excellent suggestions.

Charlie is very open to new ideas; but he is also very firm. He won't let me do anything unless he knows it works and is better than the current process. Player experience comes first. TD experience next.

In Gatlinburg, Bridgescore+ is going to be used at 9, 1, 4:45, 7:30. I'm going to be supporting Bridgescore+, and playing, most likely on a 5+ person KO team. At some point I need to eat, and sleep, and possibly have a social life. There's no time in my schedule for another meeting (I already have some). There's some new technology we're rolling out in Gatlinburg. I need to make sure it all works. Gatlinburg is a big stage to screw up on.

I'll be happy to show anyone Bridgescore+ while I'm there, and have time, but this must not interfere with the tournament operations. District 7 is looking at using this technology at all of their tournaments so educating them is my priority.
March 11, 2015
.

Bottom Home Top