So after spending many late nights creating demos, getting swag together, assembling booth material and otherwise preparing for SpeechTEK, Open Methods has just returned from New York, somewhat triumphantly IMHO. I say this having never been to a conference with Open Methods before, since I just joined the club in February of this year, but from what I hear my experienced co-workers say, this was a good one, maybe the best yet.

Now, I heard several people say that this SpeechTEK was much smaller than last year and wonder aloud about the future of SpeechTEK itself, but I say does that really matter, if the quality of what’s left is so good? Perhaps we’re now a case of a big fish in a little pond or maybe it’s just the Tipping Point (www.gladwell.com) after 4 years as a company. Whatever the reason, the buzz about Open Methods was palpable. We were inundated at the booth and not just with looky-loos or people looking for swag either. We had lots of genuine second and third level interest from a wide variety of companies and individuals who wanted to go open source and use our tool to do it. The coming months will tell us the true result.

Now, I realize we also had very popular t-shirts and everyone just had to have one. “Speech Geek” on the front and a GRXML grammar superimposed with the words “Talk nerdy to me” on the back was refreshingly laid back and irreverent for SpeechTEK I guess. You should have heard our other candidates. Maybe we should sell them on the website, since we had numerous offers for trade or sale. In addition, we had some awesome demos, showing the use of VoiceXML and IM, VoiceXML and Video and VoiceXML and Speech. Look for our demos to be on the website soon and write us for a copy of the CD with code and getting started guide before they’re all gone.

For me, personally, getting to meet others in the industry that heretofore have been virtual relationships was great. Putting a face with a name always makes things better. I connected with old cohorts and made many new contacts. I attended several good sessions that will help hone our best practices based on large studies with good data. Participating in the Advanced Dialogs invited workshop on Thursday was a highlight, simply because of all the great minds in a room at once. Getting to “geek out” like that and revel in our researchy sides is necessary from time to time. Then it’s back to reality, because I tend to be on the pragmatic side of making (preferably speech) systems work for the end-user, rather than philosophize about what might work and I got a lot of good ideas for both sides at SpeechTEK. Malcolm Gladwell’s keynote speech on Speech at the Tipping Point really got my creative juices flowing too, so don’t be surprised if I explore a few ideas with you here on my blog in the future. Oh yeah, and I also grabbed some pretty cool swag for myself as well ;).

-Simonie

Share via:

We have sprinted, walked, stumbled, fallen down, picked up, battonned hatches, and ultimately have overcome.
After weeks of late nights, code and QA we finally have a release ready.
For those of you who don’t know,
openVXML is our open source, Gui based, VoiceXML development tool.
It is 100% eclipse based and free for everyone to use, tweak, or begin their quest for world domination (hey you’ve got to start somewhere).
Its home project is the Eclipse Voice Tools Project http://www.eclipse.org/vtp

I must say that during this release process I have learned many things. Organizing an open source development effort in a services based environment is tough, priorities are constantly shifting, resources are re-purposed, and there are days you are wondering if this is really going to work or is your boss just playing some sick little joke on you.

Somethings that helped us persevere:

1. A good team goes a long way – This I already knew, but it was re-affirmed in this release, I want to throw some props out to Trip, Lonnie, and Brian for picking up the slack and pulling through.
2. Process Counts – I know all you developers out there are moaning as you read this BUT, a simple concise process is needed if your resources are shifting. It has to be something that anyone can pick up quickly. In the instance of our QA cycles there were multiple occasions where other people had to step in and help out when my main resources were pulled onto other projects. Keeping the process simple and pointed kept the ball rolling.
3. Find something to shoot for – small victories keep people motivated. Don’t overwhelm everybody with the big picture. Everyone needs to be clued in on what your trying to achieve but your have to Break it down and set goals, don’t forget to congratulate people when the goals are met. (I know this seems obvious but it’s easy to forget.)

The trials and tribulations are probably not as important as the fact that we have finally gotten somewhere.
Keep an eye on the website. The new release will be posted soon with some updated documentation, and we’ll be posting more blog entries on some of the new 3.0 functionality

– Tyler

Share via:

OpenMethods needs a new T-shirt and we are asking you (the reader) to help. We currently have 2 T-shirts that we use on a fairly regular basis (you can see pictures of those shirts below) but we are headed to SpeechTek on August 20th – 23rd and Astricon  Sept 24th-28th. We are looking for some new ideas for T-shirt designs to take to the conference. We are looking for designs or slogans that revolve around Open Source, Voice and IVR Development Tools, Voice Applications, Speech Recognition, or anything else you might just think is cool. Post your ideas as comments to this entry, we will go through all of the ideas. If your idea is selected, we will re-post the T-shirt design and send you a T-shirt (so include your size when posting.). In the event of people posting the same ideas or concepts, the award will go to the person that posted the idea first

–Tyler

Share via:

I saw I Now Pronounce You Chuck and Larry this weekend, and not only was I subjected to yet another sub-par performance by otherwise funny comedians, but I was also affronted with some not so good natured IVR ribbing. Basically, Kevin James has a bad IVR experience and afterwards tells someone the system mistook “Brooklyn” for “Schmooklyn”, or something along those lines. Unfortunately, there have always been many things to pick apart about IVRs, but it bothers me most in this form, when it isn’t even close to accurate! “Schmooklyn” would only be an option in the recognizer to mistake for Brooklyn if there is a city called Schmooklyn in the database – and I don’t think there is, though you never know with those New Yorkers. So, make fun of the poorly concatenated output, make fun of the long prompts, make fun of the inability to get to an operator even, but this bit made fun of recognition, which has arguably made the most strides in the industry over the years and the joke made wasn’t even recognizable (pardon the pun). I should say that it wasn’t recognizable to those of us in the industry at least, because it’s not the kind of recognition error you will ever get. It just doesn’t work that way. However, what matters here, and should always be paramount, is the user perception that this is indeed what happens, or at least how it makes them feel it behaves. And here I thought we’d solved a lot of the underlying problems in this arena and that it was finally making its way out to the general psyche as well, but I guess it’s just too much fun to make jokes about something everyone has unfortunately had even a distant past experience with. They wouldn’t have written in the joke at all if they thought only a few people would get it folks.

Simonie

Share via: