danbricklin.com/log

Starting September 18, 2006
MITX Fireside Chat with Craig's List CEO Jim Buckmaster, New wikiCalc demo screencast and 0.97 beta release, Full-featured wikiCalc beta release 0.96, Essay with thoughts on patent litigation, wikiCalc moving along, What I've been up to -- wikiCalc and a patent lawsuit
18Sep06-06Dec06
2006_09_18.htm
Wednesday, December 6, 2006 
MITX Fireside Chat with Craig's List CEO Jim Buckmaster [link]
Last night I attended and recorded another MITX Fireside Chat. This one was with Craig's List CEO Jim Buckmaster. The interviewer was Scott Kirsner and there was a lot of Q&A from the audience. This was at a new venue and we had a little problem or two with the sound system (there wasn't a person operating it -- it was pre-set up) but overall the recording came out pretty well. It is an hour and 3 minutes long.

You'll find the link to it on my Podcast Page and it is in the Podcast Feed.

Here are some pictures:

 
Scott Kirsner and Jim Buckmaster and some of the audience

Tuesday, November 14, 2006 
New wikiCalc demo screencast and 0.97 beta release [link]
I've just posted a new screencast of wikiCalc along with a new release. The new release basically fixes some bugs found in 0.96, one or two of which got in the way of the demo. Keep those bug reports coming!

The screencast shows the latest version. (The last one was done in June.) In addition to the normal run through of the product features shown in the last screencast, this one has a short introduction that positions the product and adds a few minutes of demonstrations of the work flow, Live View functionality, embedding on other web pages, etc. For example, it shows viewing a consolidated report, clicking on a link to a divisional report, logging in to view that page, clicking "Edit This Page" on that report, making a change, and then viewing the updated consolidated values. It also shows using a normal web page with a form to call up another page with an embedded spreadsheet that includes calculations that are based on values provided by the form.

The new 15-minute screencast (about 22MB) is: "wikiCalc 0.97 Demonstration".

To get the latest release of wikiCalc, go to the wikiCalc Beta Test home page.

Tuesday, October 31, 2006 
Full-featured wikiCalc beta release 0.96 [link]
I've just posted a new beta release of wikiCalc: Version 0.96. This is hopefully the last beta version before 1.0. It has all of the localization work (moving strings into a few easily customizable files, etc.), all of the functions, final UI, etc. All that's left is more testing, bug fixing, and online documentation.

To see a list of additions to this version and download a copy, see the wikiCalc Beta Test Home Page.

The goal I set myself for wikiCalc 1.0 was to have a "full" product that was widely useful. I defined "full" in a computational sense as meaning that it had a good set of functions. The 109+ that it now has are listed on OpenFormula.org as being the minimum useful set as proven by some other existing products.

In a visual sense, I wanted the product to be able to produce real web pages that were at a professional enough level that corporations would feel comfortable using the product for internal use and perhaps even external use. That was a real challenge which I believe that I've met. I've been taking a variety of existing pages on the web and trying to recreate them with wikiCalc. This has caused me to continually beef up the feature set. From the first release wikiCalc had cell borders. It has a color chooser that accepts optional RGB values so you can match corporate colors. The number formatting is a near-complete implementation of that in existing commercial spreadsheets. I finally bit the bullet and even added explicit cell padding settings.

As a test, I tried recreating the tables on a financial page on the IBM Corporation web site. This appears to be a webmaster-created page on ibm.com so it is a good example of a "good enough" bar to be over. It has a variety of carefully formatted cells, lots of different colors and fonts, etc.

Here's part of the original IBM page:


Here's an attempt to recreate that page with wikiCalc. The important part is the set of tables below "Financial information":


Here's a screen shot of wikiCalc editing that example (the top and left of the final page are provided by a template which I wrote using plain HTML and CSS, just like for a blog page):


Another part of being a "full" product that is widely useful was to make the product easily localizable so that it can be an international Open Source product. That has had some effect on the UI in that many things, like lists of numeric formats, are variable and need to be able to be long drop-downs. While that might not make the product as "clean" looking, hopefully it makes it useful for a wider range of situations.

So, I think I'm finally complete for 1.0 as measured by my goals for that milestone.

What's next? I really need help testing this product. I'd really appreciate it if people would test it out soon. I need testing of the functions to make sure the calculated values are correct. I need testing in a server-based environment to see how well the Edit This Page system works, how robust the security is, etc. I'm not looking for changes to the UI to make it "better" or more like Excel, etc. I'm looking to find things that are supposed to work that don't. Even just using it to recreate something you already have and letting me know if it does or doesn't work will help. I don't need you to tell me where the bug is in the code (though that's always nice) -- I just need to know what potential bugs you found.

wikiCalc has the potential to be an important product for the Open Source community as well as for IT in general. It is a complete server-based spreadsheet that runs on your own server, not only on a service provided by others. It keeps an audit trail that may be helpful to corporations that are concerned about such things (Sarbanes-Oxley?). It works like a wiki as part of a web of potentially editable pages in a collaborative environment. It is written in a popular scripting language (Perl) that makes it well suited for experimentation. It would be great to have the 1.0 release come out after a reasonable amount of final testing.

After the 1.0 release, Socialtext will be helping me run a project on Sourceforge for upgrading this product and will be providing additional developers. Having wikiCalc (as SocialCalc) evolving alongside Socialtext's open wiki should be interesting. As Google showed this morning with their announcement of the acquisition of JotSpot, wikis and spreadsheets are both part of the new suite of web-based productivity tools.

Wednesday, October 11, 2006 
Essay with thoughts on patent litigation [link]
As I wrote last month, I was an expert witness in a patent litigation case in Eastern Texas. This venue has been further publicized in an article titled "So Small a Town, So Many Patent Suits" in the New York Times. That article mentioned the case I was involved in, Hyperion v. Outlooksoft, and brings up some issues with regards to litigating in Eastern Texas.

I was an expert on defendant Outlooksoft's side. On the witness stand, I went over some background information about the state of the art prior to the patent and explained why the patent claims were invalid in view of various specific prior art programs and documentation. One of the key pieces of "prior art" I made reference to was VP-Planner, the spreadsheet/database from Paperback Software (Adam Osborne's company), and one of our witnesses was Jim Stephenson, one of the authors of the program, who explained its functionality.

Before I left Massachusetts to go to the trial, Phil Malone of the Berkman Center for Internet & Society at Harvard Law School told me that he'd be interested in hearing my reactions after going through the process. Patent litigation in Eastern Texas is something many lawyers seem to be interested in and few have direct experience with. Phil is known, among other things, for being the DOJ's lead counsel in their investigation of Microsoft. I figured that if he thought my observations would be of interest, others might, too. So, I spent some time putting together an essay. It has thoughts I came up with while sitting there during the trial, most before I knew that we had won and before I knew that this was the first time a jury there had invalidated a software patent. (I'm telling you this so you know that the successful outcome isn't completely tainting my thinking.)

The essay is "Thoughts on Patent Litigation in 2006" and is in the Writings section of my main web site. As Phil asked, these are impressions during and right after the trial. These relate to public policy and the process in what's sometimes called the "rocket docket" for its short timeframes.

A little more about something personal I went through at the end of the trial (it has no legal or public policy consequence, but blogs allow for these things):

I didn't pay careful enough attention during the long summations and charge to the jury to remember the exact wording of the form the jurors had to return to the judge with their verdict. Sitting in the back of the courtroom, unlike the lawyers I didn't have a copy with me. When the judge read the verdict, he didn't read the complete form. He just said, "Question No. 1 as to Claim 8, all products, the answer is no. As to claim -- this is the '141 patent. As to Claim 9, as to all products, the answer is no..." Great, I thought. I remembered that the first question had to do with infringement on each version of Outlooksoft's product (not my part of the case) and that "no" meant no infringement. The defendant Outlooksoft had won the most important part to the case to their business.

Then question 2, about willfulness, another "no". Another win. Then the judge said "The answer to Question No. 3, the '141 patent, as to Claim 8..." This must be my part of the case, validity. "...yes. As to Claim 9, yes. As to the '608 patent, Claim 1 is yes. Claim 4 is yes." My heart sank. Was this that they found the patents valid? Poker faces on the legal teams. No hint to me. Question 4: zero dollars -- that must be damages. Great day for Outlooksoft and their lawyers. But I felt bad. I was sure the patent claims were invalid and I thought I got my reasons across well to the jury, but I guessed that they couldn't bring themselves to invalidate a patent or that I was much less clear than I had hoped. I guessed I didn't do that good of a job even after a lot of work.

The judge dismissed the jury, telling them they could talk about the case with whomever they wanted, but that the lawyers could not approach them -- but that they could approach the lawyers. As soon as the judge left the courtroom, things were a blur as I went up to the lawyers on my side. They were busy talking with the Outlooksoft people and did not seem to want to talk to me about any details. I guessed we really didn't win my part. I wanted to hear from the jurors before they walked away why they didn't agree with my testimony. I hurried out to the entrance way. I saw one of the jurors. I stopped and looked at her questioningly. She stopped, and then smiled with a smile that I interpreted as "I'm very sorry" and then walked over to me. She stuck out her hand to shake mine. She told me how important my testimony was to their decision (which made me feel a little better), and described a bit of how they felt about the case and something about invalidating the patent. "Invalidating???" I thought. What? I turned and called out to one of the lawyers I worked with who had just walked by. "They invalidated the claims?" He said "Yes, of course." Didn't I hear that? Oh! So it was "yes...the patent claims are invalid." Yahoooo! I did do a good job. I went out and spent some time across the street talking to a few of the jurors who had stopped some of the lawyers to talk. Slowly I started feeling much better even as I stood there in the bright, hot, early afternoon sun.

What an emotional roller coaster ride I was on. Later that afternoon I found out that this was the first time an Eastern Texas jury had invalidated a patent, and that there were only a couple other wins by non-infringement with a jury there. I was flying high as I sat on the plane home a few hours later, playing lots of "up" songs on my iPod (and enjoying my "dinner" of two Powerbars because of tight connections). It's nice to be the "first" to do something hard.

Here are some photos:

 
The road to Marshall, TX, from the Shreveport, LA, airport
 
The view outside the Hampton Inn motel and behind the front desk
 
Driving through Marshall, Texas
Driving into downtown where the courthouse is located
 
The U.S. District Court House

 
"War room" setup at local office where lawyers work far into the night on the case as it evolves and paralegals produce visuals of various sorts
 
What I looked like after my testimony down the block from the courthouse still holding a copy of my expert report (yes, I wore a tie and jacket for 5 days), Jim Stephenson who also testified and is one of the authors of VP-Planner which was a key example of prior art
 
wikiCalc moving along [link]
I've reached some milestones on the wikiCalc project. I finished coding all 109 of the "Small Group" of functions listed as the basic functions that are widely implemented in even resource constrained spreadsheets by the proposed Open Document Format formula specification. (This list started as the Level 1 list which "must be supported by all implementing applications" on the www.openformula.org site.) It includes DSUM, IRR, FV, RATE, and other useful and "fun to implement" functions. (I had to dust off my old algebra skills to take the "just solve this equation for the appropriate values" definitions and turn them into actual code, as well as do lots of black box testing with Excel and OpenOffice.org Calc.) Each function has various error checks, etc., that are very specific to spreadsheets, so there is almost always more coding than just using functions built into a language. These functions, plus the few wikiCalc-specific ones I've already shipped, bring that part of the system up to Version 1.0 functionality.

I've been finishing up the moving of strings into a localization module to ease translation into other languages and should have that done within the next day or so. All that's left before 1.0 is some last minute functionality or workflow tweaks (I have two in mind that shouldn't take more than a day or so), testing and bug fixes, and documentation. The online documentation is almost all up to date (e.g., all of the new functions are listed appropriately in the Help system). The documentation I want to have in order before 1.0 relates to demonstration videos, online setup material, etc. At that point I'll "ship" version 1.0. Anybody who would be willing to help do bug testing on a release candidate before that should please let me know.

I really can't believe that the product is getting so close to being complete. Version 1.0 has always seemed so far in the future. It's a great feeling. But not for long: Soon it's back to coding and on to Version 2.0...

Monday, September 18, 2006 
What I've been up to -- wikiCalc and a patent lawsuit [link]
As most longtime readers of this blog may have noticed, I've been posting much less frequently the last few months. The two biggest reasons have been that I've been developing wikiCalc and that I've been doing some "consulting". Finally, I'm pretty much finished with that consulting job.

The consulting job was being an expert witness in a patent infringement case between two software companies. I worked on behalf of Outlooksoft, the original defendant. There was a counterclaim patent case, but in the end the trial came down to just the two patents being asserted against Outlooksoft. The trial was held in Marshall, Texas, in the United States District Court for the Eastern District of Texas in front of Judge T. John Ward. It was a jury trial. From what I've heard (such as this article in Michael C. Smith's EDTexweblog about patent litigation in Eastern Texas) juries have only found for the defendant in a patent case in Eastern Texas once in recent years (there have been 21 or so patent verdicts) and that was for non-infringement (the patent itself was not declared invalid, which is a higher legal hurdle). (I understand from others that there may have been another non-infringement verdict.) Going into the trial, juries, as I understand it, have never invalidated a patent in Eastern Texas in recent years. (Judges have, such as when Judge Ward invalidated one earlier this month according to that blog.)

The trial was held all last week. I testified on Thursday on the subject of invalidity. Friday the jury handed down a verdict of no infringement and also invalidated all the asserted claims.

Now that this trial is over, I can hopefully spend more time on wikiCalc, podcasting, and blogging. We'll see...

Archive   Home   Newer   Older
© Copyright 1999-2014 by Daniel Bricklin
All Rights Reserved.

See disclaimer on home page.