Transana's lead developer is David Woods. This page records his thoughts on his development work. Check it frequently; David updates it fairly regularly.

On the sidebar are some links to specific development-related issues.

November 4, 2008

I've started experimenting with a multiple video interface. It's just early experimenting, but the results so far are interesting.

With MPEG-1 video on Windows, I can show 3 simultaneous videos. Synchronization isn't perfect, even as a stand-alone with just 3 video frames and none of the rest of Transana's overhead. There seem to be limits to how close I can get the separate streams running. If you have audio up in all three windows, you can hear that they're slighly out of synch, but it's only a frame or two, so if you only have one audio track playing, it actually looks pretty good. Of course, we'll have to see what happens when we add in synchronizing a transcript, etc.

With MOV video on Windows, I can't even get two video streams to play decently in my stand-alone video player app. The audio is in perfect synchronization, but the video frames simply don't update adequately. One frame is jerky, but the other doesn't seem to update at all. MOV video looks like a non-starter to me.

My second attempt with QuickTime MOV video was MUCH more successful. I just rendered it as a lower bitrate with smaller images, and it works just fine.

Obviously I have a lot more exploring to do, but at the very least, this is starting to look feasible.

October 29, 2008

Release is finally out. Yeah. Now to get the notification e-mails out.

October 28, 2008

Yeah, 2.30 is still not out. I didn't get the French translation until late on the 17th, which was the day I wanted to build the new version. Then I left town for a week in Luxembourg for a conference and a consultation with the folks at the DICA Lab at the University of Luxembourg. Then I spent yesterday catching up on everything, and today I started to build the new release. Unfortunately, in testing the French translation, I noticed a bug that slipped past the testing I'd done without translations. I'm in the process of fixing that now and doing some additional non-English testing. So the release is delayed until I am sure this new bug is fixed in all the different forms it could take.

October 16, 2008

I was hoping the 2.30 release would be ready to go out yesterday or today, but it's not. I haven't gotten several translations back from translators, and I don't want to have to do an update when those translations do come in. Due to other commitments, the release will come out on October 27th or 28th now. It should be a good one, though. I really like the new features, and already I can't live without a couple of them.

September 30, 2008

The big news of the past couple of weeks is that drag-and-drop now works on the Mac version of Transana, meaning that the Windows and Mac versions are practically identical now. Different databases now remember different video root folders, if that's how your data is set up. (The small number of you who need that will be very happy.) I'm still testing the 2.30 release package, and that testing is going extremely well. I'm finding very few issues that need to be resolved.

I've sent the 9 translation files out and have gotten three of them back already. I've at least heard from 4 of the remaining 5 translators. I'm afraid we may lose the Swedish translation this time around.

As a result of all of this, I'm expecting the release to be ready pretty soon. I'm kind of busy with presentations and papers at the moment, which is important, but it slows down the testing process. I'm not sure exactly when the release will be, but it should be within the next 2 - 4 weeks if all goes well.

July 30, 2008

Many cool new features.

Multiple simultaneous transcripts. Imagine embodyinging a different analytic focus in each of several transcripts and being able to work with them independently or in coordination. And you can create clips with multiple transcripts too.

Undo and Save added to the Search interface.

Batch Episode Creation. You select a set of files or directories and Transana will create episodes for all the media files indicated, including pulling in RTF or TXT documents as transcripts.

Moving Episodes between Series.

Default filters. If you have a filter named "Default", Transana will load it automatically when calling up a keyword visualization, map, graph, or report.

Clip Merge.

I'll be done with development for the next release pretty soon, and will then start exhaustive testing before release. I don't know how we ever lived without some of these features, now that I've got them.

March 5, 2008

I've finally got the Transana 2.22 release complete and published.

To some people, this release will seem "small," without a lot of new features. Change propagation is actually pretty significant, given the way qualitative analysis tends to flow, and there are actually a fairly good number of relatively minor improvements. Whether I like it or not, there are also a lot of bug fixes in this release, and a couple of those bugs are kind of nasty.

I recommend that everyone upgrade immediately. But then, I always recommend that everyone upgrade immediately. I frequently have the experience that some change I make to the program becomes immediately essential for me (like human-readable time code data.)

March 4, 2008

Late yesterday, I discovered a killer bug. It's something that people are not very likely to actually do, but if they do it, the problems are quite serious. Basically, your database becomes corrupted and you'll need to send it to me for repair. (Fortunately, now that I understand the problem, it's pretty easy to fix.) I may have seen the results of this bug once, but it was long enough ago, I can't be sure.

If you tell Transana to delete a Collection, it will warn you if that would cause a Keyword Example to be deleted and give you the opportunity to cancel the delete. But if you choose to interrupt the deletion process, Transana was deleting the collection without deleting the contained clips, notes, or clip keyword records. Because that leaves the database in a state where there are clips in the database that belong in a collection that's been deleted, Transana can't build the database tree correctly. Your data is no longer accessible without my intervention. (I just figure out the number of the collection that was deleted and create a collection manually with that collection number.)

This bug has probably been around for a fairly long time. However, people rarely delete whole collections, they don't use Keyword Examples much, and they very, very rarely decide to cancel a collection deletion because of the presence of a keyword example.

The source code has now been fixed to avoid this problem and the fix will be included in the 2.22 release. My apologies to anyone who may have been affected.

February 29, 2008

I'm on page 24 of the test plan, the last page. The things I have left to test are things that shouldn't have changed, with one small exception. I'm just waiting for the last couple of translations to come in. The new release, version 2.22, should be available on Monday or Tuesday of next week.

The release after that will be made up of changes for two projects. Erica Halverson and her Youth Filmmaking Project at the University of Wisconsin - Madison need multiple simultaneous transcripts. The plan is to allow up to four transcripts to be opened for a single video file. And Nicolas Sheon of the Center for Aids Prevention Studies at the University of California, San Francisco has requested a number of improvements to help facilitate his work. I think both of these projects will prove beneficial to many Transana users around the world.

January 30, 2008

So close, and yet so far.

I was doing some final testing for the 2.22 release when I found a bug on both of the Mac versions. My current theory is that this bug comes from wxPython or somewhere upstream from that, as I've been able to demonstrate the problem outside of Transana on two different computers with different setups.

I've reported the problem, and will continue testing for the next couple of weeks. Either a fix will come in, or the next release won't have Print Preview on the Mac. I don't see this as a serious problem, since the new report infrastructure greatly reduced the necessity for Print Preview anyway.

November 16, 2007

Where does the time go? In the past 5 weeks, I've put out the new release, created a Demonstration version of Transana, discovered and fixed a few bugs, and added a couple of minor new features. For example, in the summary of the Episode and Collection Reports, I've added cumulative time for each keyword to where keyword counts are displayed.

Right now I'm working on Transcript Change Propogation. When you change an Episode transcript, you will be able to broadcast those changes out to clips that have been created from that transcript. (But you don't have to update the clips where it's not analytically appropriate to do so.)

There have been a number of people who have asked for this sort of functionality over about the past year. But keeping clip transcripts and the source episode transcript synchronized turns out to be a lot harder to do properly than people seem to think. What I'm building is more subtle and deliberate than what people have asked for, but I think it needs to be. Special thanks to Alain Desilets and Jen Patashnick for their input.

October 8, 2007

I built the Transana 2.21-MU Beta 1.4 release set this morning. One more trip through the test plan and hopefully I'll be able to build the final release.

September 28, 2007

I'm still testing, finding bugs, making new Beta builds, and testing some more. It's a slow, painful, unavoidable process. I'm currently running three copies of Transana-MU, one in Russian on a PPC-Mac, one in Spanish on Windows XP, and one in German on an Intel-Mac. I'm doing my best to be slow, thorough, and methodical. I'm finding enough that it's worth doing. It's almost all Unicode stuff.

I've heard developers of other programs brag that they do really well at creating development deadlines and meeting them, allowing them to get releases out "on time." Then I hear that their products are buggy and require "service packs" to work right.

I'm taking a rather different approach with Transana. Instead of setting and trying to meet artificial deadlines, I'm testing the software until Ifeel confident that things work well, and then I'll do a release. Of course there will be bugs in Transana releases. There are bugs in all software. And when people report them, I will fix them. (And I usually find more bugs than get reported to me from the field.) But I think my approach leads to more stable software on a release-by-release basis. It also means that it's very hard for me to predict when the next release will be.

September 14, 2007

I've been out of town a LOT lately. I ended up taking almost 3 weeks off in August, and another week off in early September. I'm now rested up and my oldest son is safely delivered to college to start his freshman year.

It turns out that there's nothing like being out of town with limited internet access to bring out posts to the discussion board. There have been a whole host of unicode and internationalization issues with Transana 2.20, particularly on the Intel-Mac. You couldn't have non-English characters in database names on Intel Macs. You couldn't end keywords with the "à" character. You couldn't export and import databases with notes that used non-English characters. These issues have all been fixed, as far as I can tell.

There's also been this persistent issue on the Intel-Mac with certain error dialogs showing up with no text, just an "OK" button. (If you see this, just hit the "Tab" key and the contents of the dialog box will magically appear!) I have not been able to reproduce this in a simple example, which suggests there may be an interaction between several factors contributing, perhaps including internationalization routines, unicode characters, and the wxPython wxMessageDialog. I've started replacing the wxMessageDialog with an identical dialog I wrote myself, and that seems to be working.

So I've now got a good Beta build that I'm testing. And I'm going to be around for a while. So the new release should be ready within the next few weeks if all goes well.

August 2, 2007

Still no Beta. I've found a couple of categories of bugs I'm fighting with. It's all Unicode stuff I missed with my first-round testing in English. Some of it is a bit tricky, requiring two copies of Transana-MU working in different languages. Some of it just requires data with non-English characters. Sigh. The Beta will not be available until I resolve these issues.

July 20, 2007

I decided recently not to wait until the Research Journals were built before putting out the next release. I realized that Research Journals are going to be a larger project than I'd given them credit for, so they will be one of the major innovations for release after next. And with the new report infrastructure and the Notes Browser, this release is already going to be pretty important from an analytic standpoint.

I've been testing the new release all week. I'm currently on page 13 of my 21 page test plan. It's going well, and as expected, I'm finding bugs. Many of the bugs are new, ones that are part of the new features or are the byproduct of code changes brought about by implementing the new features. One I found today was something that someone had reported several months ago but I was never able to recreate before. A couple have been around for a while. (Did you know you can't zoom in on the Visualization if your selection starts at 0.0? I'm guessing that's been a problem since Transana 2.0.)

I'm hoping to have a Beta test version by August 1st. I'm planning to send the new prompts to the translators by that time. That would put me on schedule for a September release. In the past, I've sent the prompts to the translators earlier in the process, but that meant they couldn't see the features they were translating, and I often found bugs requiring changes to prompts during early testing, which meant that translation took two rounds. I'm also not planning on releasing the Beta publicly, as I've done in the past. I rarely got much feedback from the Beta tests anyway. If you want to be part of the Beta test, send me an e-mail.

July 3, 2007

Okay, the Notes Browser now appears finished. I still need to do quite a bit of testing with it, but it's looking pretty good to me at the moment. Things are moving slowly because I've just got too much going on right now, what with conferences and vacation travel. I'm going to spend the next couple of weeks testing and trying up loose ends before I start on the research journals. I still don't have any idea when the next release will be.

June 8, 2007

The new report infrastructure is mostly done. I added a bunch of customization options to it so you have a lot of control over what is included. I've expanded Clip Data Export to include an "All Clips from All Episodes in a Series" and "All Clips" options. I added "Next" and "Previous" buttons to Play All Clips, as well as making it aware of clips in nested collections.

I've improved the Notes Editor quite a bit, so you can now export notes as *txt files and print them. I also added a button for inserting date and time, as well as adding text search within a note. I'm currently working on a Notes Browser too. The idea behind this is that Notes are (or at least can be) documentation of your analytic process, and they weren't easy enough to scan through to find important ideas that were recorded in the notes. So the Notes Browser allows you to quickly find and look through your notes, as well as being the place where you produce Notes Reports.

Also, this week I added the ability to import plain-text transcripts. I did this at the request of Brian MacWhinney, so he and his colleagues could import CLAN transcript files into Transana for additional analysis. At the moment, this only works for transcripts in English.

Next week, I hope to finish up the Notes Browser, including adding the capacity to search for text across notes. After that, I hope to move on to Research Journals, which are another tool for documenting the analytic process and writing about the results of your work in Transana.

May 1, 2007

April sales of Transana were ... less than I'd hoped for, given the features and quality of the new release. There were about 6 times the number of downloads of the old free version as there were purchases. Sure, I understand free is better than pay, but people need to understand that if we don't generate adequate sales, Transana development will cease. That's the cold economic reality. So please, make your Transana purchase today, and support the program's ongoing development.

I'm pretty pleased with the way the new reports are coming together. I should be able to finish up the first round of improvements by the end of the week.

I've got the Keyword Summary Report done, and the Collection Summary Report and the Keyword Usage Reports are almost done. (I just have to make sure that all the changes I made work with the search-results versions of the reports, which shouldn't be hard.

In addition to all the features listed in my last post, there are a couple of other features I'd like to mention. First, where Collection names are show, I've made an effort to show the Collection's full nesting information. Second, I've made it so that you can (optionally) include clips from nested collections in the reports. It used to be that the reports could only look at clips in the selected collection, but of course it makes sense to be able to include the clips in nested collections in the reports, since they are supposed to be part of a sub-categorization. So that is now an option. (In fact, it's the default.) Finally, it became clear to me that the Collection Summary Report and the Collection Keyword Usage Report had a LOT of overlap, so I've refactored the report infrastructure so they're actually produced by the same code. With the new options available, you can actually produce both reports from either menu option, including a version that includes both transcript data and keyword data for the same set of Clips.

The report infrastructure is very flexible, so I could expand the reports to be more inclusive and customizable. I'm just not sure what would really be useful, what else would be worth my time to implement. I could add an option to include Title/Comment information from the data objects, but does anybody really use those fields? I'm open to suggestions for reports you think would be helpful. Just e-mail me with your ideas.

April 27, 2007

The new report infrastructure is just about finished and the first report, the Keyword Summary Report, is just about finished. (Yes, I picked that one because I knew it would be the easiest to convert.)

Advantages:

• The report is customizable. It has a filter dialog so you can select what keyword groups to include or exclude from the report. And I can pretty easily add more options if there's a need.
  
• The report is editable. You can work with it like any text document, adding, changing, and deleting text at will.
  
• You can copy and paste selected text from the report to a word-processing document for inclusion in articles. (Some formatting will be required.)
  
• You can easily save the report as a Rich Text Format (RTF) document.
  
• You can print the customized and edited report.

Disadvantages:

• The new version of the report isn't as pretty as the old version.

Overall, I'm pretty happy with it, and I think it will meet most of the needs I've heard from researchers that the old report system didn't. An added bonus is that the new infrastructure is easier to work with, making the creation of additional text reports by others through the Open Source model easier.

April 18, 2007

I've started building some infrastructure to revise how the text reports are created. The idea is to make it so that you can customize, edit, and save the text reports to RTF files, as well as printing them like you can now.

It feels good to code again. It's been a while since I did any heavy lifting with the code. But that always happens around a big release.

April 3, 2007

Transana 2.20 was released today. We finally have a full-functioning version for the Mac. And the new version includes many cool features I've been working on for a long time.

March 29, 2007

We have no choice but to start charging for copies of Transana. Funds available through targeted development, donations, and grants are not adequate at this time to ensure Transana's ongoing development, so we chose to start charging rather than to reduce the amount of time being devoted to developing and supporting the program.