ClearTK 2.0 release

[Philip Ogren]

Here's my straw man suggestion to get this release out the door: 

- Move all the issues marked as 2.1 to 2.2.  

- Move all the issues marked as 2.0 to 2.1 except 388 and 394

- Fix 388 and 394

- Release 2.0

How does that sound?

[Lee Becker]

I'm a bit rusty, but most of the API changes have been checked in already, right?  I also think we should just bump the maven dependency on ClearNLP.  Jinho released a minor revision update recently.

Otherwise sounds like a plan to me.

[Steven Bethard]

On Sun, Feb 9, 2014 at 9:58 PM, Philip Ogren <phi...@ogren.info> wrote:
> Here's my straw man suggestion to get this release out the door:
>
> - Move all the issues marked as 2.1 to 2.2.

Why is this necessary? Why not just leave them as 2.1?

> - Move all the issues marked as 2.0 to 2.1 except 388 and 394
> - Fix 388 and 394

So, specifically, you're suggesting we move to 2.1:

set up continuous integration server
https://code.google.com/p/cleartk/issues/detail?id=197

Tutorial on viterbi
https://code.google.com/p/cleartk/issues/detail?id=283

POS tagger work
https://code.google.com/p/cleartk/issues/detail?id=364

unit test trainable feature extractors
https://code.google.com/p/cleartk/issues/detail?id=378

consider switching to UIMA resources for classifiers, etc.
https://code.google.com/p/cleartk/issues/detail?id=393

I'm okay with postponing all of these.

> - Release 2.0
>
> How does that sound?

There are several issues at the end that at least need to be assigned
a milestone:

XmiWriter in examples should be moved to cleartk-util
https://code.google.com/p/cleartk/issues/detail?id=391

LineReader creates a non-file URI
https://code.google.com/p/cleartk/issues/detail?id=392

Losing cause exception stack traces
https://code.google.com/p/cleartk/issues/detail?id=398

Division by zero in ZeroMeanUnitStddevExtractor
https://code.google.com/p/cleartk/issues/detail?id=399

I think the last two at least need to be fixed before 2.0 since these
are clear bugs and have been reported by an outside user, and we'd
like to encourage that as much as possible.

Steve