Java version in Tika

classic Classic list List threaded Threaded
12 messages Options
Reply | Threaded
Open this post in threaded view
|

Java version in Tika

Jukka Zitting
Hi,

I noticed that both the Nutch metadata and the Lius lite code are
using Java 5 features.

Do you think we should require Java 5 in Tika, or should we support
also Java 1.4?

I believe there are a number of potential users (Jackrabbit included)
that are still bound to Java 1.4, and since Java 1.4 will not reach
end of life status any time soon (there isn't even a JSR for Java 7
yet) I personally think we should stick to 1.4 and remove the Java 5
features from Tika.

BR,

Jukka Zitting
Reply | Threaded
Open this post in threaded view
|

Re: Java version in Tika

Bertrand Delacretaz
On 6/13/07, Jukka Zitting <[hidden email]> wrote:

> ...Do you think we should require Java 5 in Tika, or should we support
> also Java 1.4?...

I think it would be wise for the Tika *framework* to work with Java 1.4.

Specific plugins can require higher versions, but requiring 1.5 for
the framework might be too limiting at the moment.

-Bertrand
Reply | Threaded
Open this post in threaded view
|

Re: Java version in Tika

Thilo Goetz
Bertrand Delacretaz wrote:

> On 6/13/07, Jukka Zitting <[hidden email]> wrote:
>
>> ...Do you think we should require Java 5 in Tika, or should we support
>> also Java 1.4?...
>
> I think it would be wise for the Tika *framework* to work with Java 1.4.
>
> Specific plugins can require higher versions, but requiring 1.5 for
> the framework might be too limiting at the moment.
>
> -Bertrand

I agree.  UIMA is also still on Java 1.4, and I think we'll probably stay
there at least for the rest of the year.  So it would be good if the basic
framework could work with 1.4.

--Thilo
Reply | Threaded
Open this post in threaded view
|

Re: Java version in Tika

chrismattmann
In reply to this post by Jukka Zitting
Hi Guys,

 I don't have a strong preference for this either way. I simply ported over
the code for the Nutch metadata framework, which happened to use some
features from 1.5. I was a proponent of staying with 1.4 in the Nutch
community, however, was overruled there ;)

 Just to stir the pot a bit, I will say though, that IMHO, Java 1.5 is
entirely stable, has a strong, solid user base, and is becoming pervasive.
It also offers a lot of features that I think are truly useful. Is it too
much to ask of projects like Jackrabbit, etc., for them to create JDK 1.5
builds? I know that this is possible even if the project only uses JDK 1.4
language constructs, right?

 I think we may have to call for a vote on this though. What do others
think?

Thanks!

Cheers,
  Chris



On 6/13/07 3:43 AM, "Jukka Zitting" <[hidden email]> wrote:

> Hi,
>
> I noticed that both the Nutch metadata and the Lius lite code are
> using Java 5 features.
>
> Do you think we should require Java 5 in Tika, or should we support
> also Java 1.4?
>
> I believe there are a number of potential users (Jackrabbit included)
> that are still bound to Java 1.4, and since Java 1.4 will not reach
> end of life status any time soon (there isn't even a JSR for Java 7
> yet) I personally think we should stick to 1.4 and remove the Java 5
> features from Tika.
>
> BR,
>
> Jukka Zitting


Reply | Threaded
Open this post in threaded view
|

Re: Java version in Tika

Thilo Goetz
Hi Chris,

the issue is not the project itself, e.g., Jackrabbit, but the applications
that build on it.  There are many products out in commercial land that are
still tied to Java 1.4.  Once you start prereqing Java 1.5, those products
will no longer be able to use your component.  So the lower down in the food
chain you are, the more you need to cater for backwards versions of Java
(at least if you're interested in being used by commercial products).

--Thilo

Chris Mattmann wrote:

> Hi Guys,
>
>  I don't have a strong preference for this either way. I simply ported over
> the code for the Nutch metadata framework, which happened to use some
> features from 1.5. I was a proponent of staying with 1.4 in the Nutch
> community, however, was overruled there ;)
>
>  Just to stir the pot a bit, I will say though, that IMHO, Java 1.5 is
> entirely stable, has a strong, solid user base, and is becoming pervasive.
> It also offers a lot of features that I think are truly useful. Is it too
> much to ask of projects like Jackrabbit, etc., for them to create JDK 1.5
> builds? I know that this is possible even if the project only uses JDK 1.4
> language constructs, right?
>
>  I think we may have to call for a vote on this though. What do others
> think?
>
> Thanks!
>
> Cheers,
>   Chris
>
>
>
> On 6/13/07 3:43 AM, "Jukka Zitting" <[hidden email]> wrote:
>
>> Hi,
>>
>> I noticed that both the Nutch metadata and the Lius lite code are
>> using Java 5 features.
>>
>> Do you think we should require Java 5 in Tika, or should we support
>> also Java 1.4?
>>
>> I believe there are a number of potential users (Jackrabbit included)
>> that are still bound to Java 1.4, and since Java 1.4 will not reach
>> end of life status any time soon (there isn't even a JSR for Java 7
>> yet) I personally think we should stick to 1.4 and remove the Java 5
>> features from Tika.
>>
>> BR,
>>
>> Jukka Zitting
>
Reply | Threaded
Open this post in threaded view
|

Re: Java version in Tika

Bertrand Delacretaz
On 6/13/07, Thilo Goetz <[hidden email]> wrote:

> ...the issue is not the project itself, e.g., Jackrabbit, but the applications
> that build on it.  There are many products out in commercial land that are
> still tied to Java 1.4...

Exactly...I don't think the Tika framework part will be complex enough
to justify requiring 1.5 now, and it could close the door for quite a
few projects and users.

-Bertrand (starting to hate having to go back to 1.4 here and there,
but that's life ;-)
Reply | Threaded
Open this post in threaded view
|

Re: Java version in Tika

Doug Cutting
In reply to this post by Thilo Goetz
Thilo Goetz wrote:
> the issue is not the project itself, e.g., Jackrabbit, but the applications
> that build on it.  There are many products out in commercial land that are
> still tied to Java 1.4.  Once you start prereqing Java 1.5, those products
> will no longer be able to use your component.

That's true.  But Tika is still gestating.  We should pick the highest
Java version that we think most clients will be moving to over the next
few years, as Tika matures and is adopted by projects.  So, if we assume
that Tika will be mature and ready-to-use in a year, and that it will
then be incorporated in development versions of applications, that might
themselves not then released for a year, then the question is not
whether these applications are using Java 1.5 today, but whether these
applications will be using Java 1.5 in two years.

Different applications will incorporate Tika at different times and will
upgrade to Java 1.5 at different times.  Choosing 1.5 will undoubtedly
rule out certain applications (some applications may *never* upgrade to
Java 1.5).  But I'd argue that the applications that are most likely to
be interested in upgrading to the latest-and-greatest Tika-based tools
are likely to be those that will also be upgrading to Java 1.5.

Platform availability used to be an issue for Java 1.5.  It isn't much
any longer.  IBM, GCJ and Apache Harmony all support Java 1.5 features.

Doug
Reply | Threaded
Open this post in threaded view
|

Re: Java version in Tika

Keith R. Bennett
In reply to this post by Thilo Goetz
All -

My preference would be to go to 1.5, to use generics, simpler loop syntax, and other new features, but it sounds like the concensus is to stay compatible with 1.4.  If that is the case, would it be helpful for me to enter a JIRA issue for this?  And make the source code changes?

- Keith Bennett


Thilo Goetz wrote
Bertrand Delacretaz wrote:
> On 6/13/07, Jukka Zitting <jukka.zitting@gmail.com> wrote:
>
>> ...Do you think we should require Java 5 in Tika, or should we support
>> also Java 1.4?...
>
> I think it would be wise for the Tika *framework* to work with Java 1.4.
>
> Specific plugins can require higher versions, but requiring 1.5 for
> the framework might be too limiting at the moment.
>
> -Bertrand

I agree.  UIMA is also still on Java 1.4, and I think we'll probably stay
there at least for the rest of the year.  So it would be good if the basic
framework could work with 1.4.

--Thilo
Reply | Threaded
Open this post in threaded view
|

Re: Java version in Tika

chrismattmann
Hi Keith,

 I think at this point, it would be a waste of your time to do the work to
convert the code to 1.5. Since it's really a project decision, when the
project is ready to move to 1.5 (which it doesn't seem like it is yet), then
it would probably make sense at that time to do it, rather than to do it
now, and watch your patch widdle away and become out of date with each code
change. So, IMO I wouldn't worry about patching the code to use 1.5...yet :)

 As for the JIRA issue, I think it would be a good idea to track (say, as a
JIRA "task") the desire to move the code to 1.5. Then we (the developers and
community) could vote on the issue, and track it.

 Thanks!

Cheers,
  Chris



On 9/10/07 9:36 AM, "kbennett" <[hidden email]> wrote:

>
> All -
>
> My preference would be to go to 1.5, to use generics, simpler loop syntax,
> and other new features, but it sounds like the concensus is to stay
> compatible with 1.4.  If that is the case, would it be helpful for me to
> enter a JIRA issue for this?  And make the source code changes?
>
> - Keith Bennett
>
>
>
> Thilo Goetz wrote:
>>
>> Bertrand Delacretaz wrote:
>>> On 6/13/07, Jukka Zitting <[hidden email]> wrote:
>>>
>>>> ...Do you think we should require Java 5 in Tika, or should we support
>>>> also Java 1.4?...
>>>
>>> I think it would be wise for the Tika *framework* to work with Java 1.4.
>>>
>>> Specific plugins can require higher versions, but requiring 1.5 for
>>> the framework might be too limiting at the moment.
>>>
>>> -Bertrand
>>
>> I agree.  UIMA is also still on Java 1.4, and I think we'll probably stay
>> there at least for the rest of the year.  So it would be good if the basic
>> framework could work with 1.4.
>>
>> --Thilo
>>
>>

______________________________________________
Chris Mattmann, Ph.D.
[hidden email]
Key Staff Member
Modeling and Data Management Systems Section (387)
Data Management Systems and Technologies Group

_________________________________________________
Jet Propulsion Laboratory            Pasadena, CA
Office: 171-266B                        Mailstop:  171-246
_______________________________________________________

Disclaimer:  The opinions presented within are my own and do not reflect
those of either NASA, JPL, or the California Institute of Technology.


Reply | Threaded
Open this post in threaded view
|

Re: Java version in Tika

Keith R. Bennett
Chris -

Thanks for the speedy reply.

There was a little miscommunication.  I meant to enter a JIRA issue to switch to 1.4, not 1.5.  If I'm not mistaken, there are 15-20 source files that contain collections with type specifications, and will therefore not compile in 1.4.)

Although it will pain me to remove the type specifications ;), I volunteer to do so for the greater good.

- Keith


Chris Mattmann wrote
Hi Keith,

 I think at this point, it would be a waste of your time to do the work to
convert the code to 1.5. Since it's really a project decision, when the
project is ready to move to 1.5 (which it doesn't seem like it is yet), then
it would probably make sense at that time to do it, rather than to do it
now, and watch your patch widdle away and become out of date with each code
change. So, IMO I wouldn't worry about patching the code to use 1.5...yet :)

 As for the JIRA issue, I think it would be a good idea to track (say, as a
JIRA "task") the desire to move the code to 1.5. Then we (the developers and
community) could vote on the issue, and track it.

 Thanks!

Cheers,
  Chris



On 9/10/07 9:36 AM, "kbennett" <kbennett@bbsinc.biz> wrote:

>
> All -
>
> My preference would be to go to 1.5, to use generics, simpler loop syntax,
> and other new features, but it sounds like the concensus is to stay
> compatible with 1.4.  If that is the case, would it be helpful for me to
> enter a JIRA issue for this?  And make the source code changes?
>
> - Keith Bennett
>
>
>
> Thilo Goetz wrote:
>>
>> Bertrand Delacretaz wrote:
>>> On 6/13/07, Jukka Zitting <jukka.zitting@gmail.com> wrote:
>>>
>>>> ...Do you think we should require Java 5 in Tika, or should we support
>>>> also Java 1.4?...
>>>
>>> I think it would be wise for the Tika *framework* to work with Java 1.4.
>>>
>>> Specific plugins can require higher versions, but requiring 1.5 for
>>> the framework might be too limiting at the moment.
>>>
>>> -Bertrand
>>
>> I agree.  UIMA is also still on Java 1.4, and I think we'll probably stay
>> there at least for the rest of the year.  So it would be good if the basic
>> framework could work with 1.4.
>>
>> --Thilo
>>
>>

______________________________________________
Chris Mattmann, Ph.D.
Chris.Mattmann@jpl.nasa.gov
Key Staff Member
Modeling and Data Management Systems Section (387)
Data Management Systems and Technologies Group

_________________________________________________
Jet Propulsion Laboratory            Pasadena, CA
Office: 171-266B                        Mailstop:  171-246
_______________________________________________________

Disclaimer:  The opinions presented within are my own and do not reflect
those of either NASA, JPL, or the California Institute of Technology.

Reply | Threaded
Open this post in threaded view
|

Re: Java version in Tika

chrismattmann
Hi Keith,

 Got it: yep, I would say, please patch the 1.5 code to use 1.4 and create a
JIRA issue for it. I'll take a look at it as soon as you put it up and we'll
try to get it shepherded into the sources as quickly as possible. Thanks for
your interest in the project!

Cheers,
  Chris



On 9/10/07 9:57 AM, "kbennett" <[hidden email]> wrote:

>
> Chris -
>
> Thanks for the speedy reply.
>
> There was a little miscommunication.  I meant to enter a JIRA issue to
> switch to 1.4, not 1.5.  If I'm not mistaken, there are 15-20 source files
> that contain collections with type specifications, and will therefore not
> compile in 1.4.)
>
> Although it will pain me to remove the type specifications ;), I volunteer
> to do so for the greater good.
>
> - Keith
>
>
>
> Chris Mattmann wrote:
>>
>> Hi Keith,
>>
>>  I think at this point, it would be a waste of your time to do the work to
>> convert the code to 1.5. Since it's really a project decision, when the
>> project is ready to move to 1.5 (which it doesn't seem like it is yet),
>> then
>> it would probably make sense at that time to do it, rather than to do it
>> now, and watch your patch widdle away and become out of date with each
>> code
>> change. So, IMO I wouldn't worry about patching the code to use 1.5...yet
>> :)
>>
>>  As for the JIRA issue, I think it would be a good idea to track (say, as
>> a
>> JIRA "task") the desire to move the code to 1.5. Then we (the developers
>> and
>> community) could vote on the issue, and track it.
>>
>>  Thanks!
>>
>> Cheers,
>>   Chris
>>
>>
>>
>> On 9/10/07 9:36 AM, "kbennett" <[hidden email]> wrote:
>>
>>>
>>> All -
>>>
>>> My preference would be to go to 1.5, to use generics, simpler loop
>>> syntax,
>>> and other new features, but it sounds like the concensus is to stay
>>> compatible with 1.4.  If that is the case, would it be helpful for me to
>>> enter a JIRA issue for this?  And make the source code changes?
>>>
>>> - Keith Bennett
>>>
>>>
>>>
>>> Thilo Goetz wrote:
>>>>
>>>> Bertrand Delacretaz wrote:
>>>>> On 6/13/07, Jukka Zitting <[hidden email]> wrote:
>>>>>
>>>>>> ...Do you think we should require Java 5 in Tika, or should we support
>>>>>> also Java 1.4?...
>>>>>
>>>>> I think it would be wise for the Tika *framework* to work with Java
>>>>> 1.4.
>>>>>
>>>>> Specific plugins can require higher versions, but requiring 1.5 for
>>>>> the framework might be too limiting at the moment.
>>>>>
>>>>> -Bertrand
>>>>
>>>> I agree.  UIMA is also still on Java 1.4, and I think we'll probably
>>>> stay
>>>> there at least for the rest of the year.  So it would be good if the
>>>> basic
>>>> framework could work with 1.4.
>>>>
>>>> --Thilo
>>>>
>>>>
>>
>> ______________________________________________
>> Chris Mattmann, Ph.D.
>> [hidden email]
>> Key Staff Member
>> Modeling and Data Management Systems Section (387)
>> Data Management Systems and Technologies Group
>>
>> _________________________________________________
>> Jet Propulsion Laboratory            Pasadena, CA
>> Office: 171-266B                        Mailstop:  171-246
>> _______________________________________________________
>>
>> Disclaimer:  The opinions presented within are my own and do not reflect
>> those of either NASA, JPL, or the California Institute of Technology.
>>
>>
>>
>>

______________________________________________
Chris Mattmann, Ph.D.
[hidden email]
Key Staff Member
Modeling and Data Management Systems Section (387)
Data Management Systems and Technologies Group

_________________________________________________
Jet Propulsion Laboratory            Pasadena, CA
Office: 171-266B                        Mailstop:  171-246
_______________________________________________________

Disclaimer:  The opinions presented within are my own and do not reflect
those of either NASA, JPL, or the California Institute of Technology.


Reply | Threaded
Open this post in threaded view
|

Re: Java version in Tika

Keith R. Bennett
Chris and All -

I've added a JIRA issue to downgrade the code to 1.4 level.  It's interesting that until now the 1.5 features code didn't seem to stop anyone from developing with it.  Is this possibly a sign?  Should we defer downgrading the code until absolutely necessary, in the hope that that time will never come?

- Keith

Chris Mattmann wrote
Hi Keith,

 Got it: yep, I would say, please patch the 1.5 code to use 1.4 and create a
JIRA issue for it. I'll take a look at it as soon as you put it up and we'll
try to get it shepherded into the sources as quickly as possible. Thanks for
your interest in the project!

Cheers,
  Chris