You are viewing a plain text version of this content. The canonical link for it is here.
Posted to server-dev@james.apache.org by Robert Burrell Donkin <ro...@gmail.com> on 2008/11/18 20:42:38 UTC

[mailet] Creaet Bayesian-Mailet product? [WAS Re: [jira] Commented: (JAMES-621) Move BayesianAnalyzer to extra project]

opinions?

- robert

On Tue, Nov 18, 2008 at 5:59 PM, Robert Burrell Donkin (JIRA)
<se...@james.apache.org> wrote:
>
>    [ https://issues.apache.org/jira/browse/JAMES-621?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12648676#action_12648676 ]
>
> Robert Burrell Donkin commented on JAMES-621:
> ---------------------------------------------
>
> We could break out the analzers and mailets out into a separate product under mailets without too much difficulty.
>
> We would probably need to fork SQLUtils (but it's a simple class).
>
>> Move BayesianAnalyzer to extra project
>> --------------------------------------
>>
>>                 Key: JAMES-621
>>                 URL: https://issues.apache.org/jira/browse/JAMES-621
>>             Project: JAMES Server
>>          Issue Type: Wish
>>            Reporter: Norman Maurer
>>            Priority: Minor
>>
>> It whould be cool to move the BayesianAnalyzer stuff to an extra subproject. This whould make it easy for other developers to use our code.
>
> --
> This message is automatically generated by JIRA.
> -
> You can reply to this email to add a comment to the issue online.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
> For additional commands, e-mail: server-dev-help@james.apache.org
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
For additional commands, e-mail: server-dev-help@james.apache.org


Re: [mailet] Creaet Bayesian-Mailet product? [WAS Re: [jira] Commented: (JAMES-621) Move BayesianAnalyzer to extra project]

Posted by Norman Maurer <no...@apache.org>.
Hi,

I still think it would be good to factor it out.. there aren't enought good
java based spamfilter libraries out there...
I think many developers would us our work if it would be possible and
visible to them.

Cheers,
Norman


2008/11/21 Robert Burrell Donkin <ro...@gmail.com>

> On Wed, Nov 19, 2008 at 1:09 PM, Robert Burrell Donkin
> <ro...@gmail.com> wrote:
> > On Wed, Nov 19, 2008 at 9:30 AM, Stefano Bagnara <ap...@bago.org>
> wrote:
> >> Robert Burrell Donkin ha scritto:
> >>> opinions?
> >>
> >> Unless there is someone willing to work on that code I think it is bad
> >> to move it out.
> >
> > the code has very little visibility within James. a mini-site would
> > have more visibility.
> >
> > i agree with norman's point that the bayesian code is potentially
> > reusable if it were in a separate library
> >
> >> Otherwise we'll end up with a microlibrary for the ClamAVScanner, for
> >> the SpamAssassinInvoker, for the NetworkMatcher and so on.
> >
> > i like micro-libraries :-)
> >
> > they've also proved very successful over the years
>
> on balance, it's important to choose the right granularity. probably
> better to pick a cohesive topic. i'm very interested in using machine
> learning with mail. if anyone starts to be become interested in
> developing any more machine learning mailets, we can factor it out at
> that stage.
>
> - robert
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
> For additional commands, e-mail: server-dev-help@james.apache.org
>
>

Re: [mailet] Creaet Bayesian-Mailet product? [WAS Re: [jira] Commented: (JAMES-621) Move BayesianAnalyzer to extra project]

Posted by Robert Burrell Donkin <ro...@gmail.com>.
On Wed, Nov 19, 2008 at 1:09 PM, Robert Burrell Donkin
<ro...@gmail.com> wrote:
> On Wed, Nov 19, 2008 at 9:30 AM, Stefano Bagnara <ap...@bago.org> wrote:
>> Robert Burrell Donkin ha scritto:
>>> opinions?
>>
>> Unless there is someone willing to work on that code I think it is bad
>> to move it out.
>
> the code has very little visibility within James. a mini-site would
> have more visibility.
>
> i agree with norman's point that the bayesian code is potentially
> reusable if it were in a separate library
>
>> Otherwise we'll end up with a microlibrary for the ClamAVScanner, for
>> the SpamAssassinInvoker, for the NetworkMatcher and so on.
>
> i like micro-libraries :-)
>
> they've also proved very successful over the years

on balance, it's important to choose the right granularity. probably
better to pick a cohesive topic. i'm very interested in using machine
learning with mail. if anyone starts to be become interested in
developing any more machine learning mailets, we can factor it out at
that stage.

- robert

---------------------------------------------------------------------
To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
For additional commands, e-mail: server-dev-help@james.apache.org


Re: [mailet] Creaet Bayesian-Mailet product? [WAS Re: [jira] Commented: (JAMES-621) Move BayesianAnalyzer to extra project]

Posted by Robert Burrell Donkin <ro...@gmail.com>.
On Wed, Nov 19, 2008 at 9:30 AM, Stefano Bagnara <ap...@bago.org> wrote:
> Robert Burrell Donkin ha scritto:
>> opinions?
>
> Unless there is someone willing to work on that code I think it is bad
> to move it out.

the code has very little visibility within James. a mini-site would
have more visibility.

i agree with norman's point that the bayesian code is potentially
reusable if it were in a separate library

> Otherwise we'll end up with a microlibrary for the ClamAVScanner, for
> the SpamAssassinInvoker, for the NetworkMatcher and so on.

i like micro-libraries :-)

they've also proved very successful over the years

> You proved that you take care of releasing what you proposed to extract,
> so I believe in you if you want to follow this path.

there are some interesting stuff which could be done using machine
learning and mail. i can think of a few cool SoC projects based around
this.

but i agree that the effect is only worthwhile if there are people
interested in this area of james

> We have
> org.apache.james.util.BayesianAnalyzer
> org.apache.james.util.JDBCBayesianAnalyzer
> org.apache.james.transport.mailets.BayesianAnalysis
> org.apache.james.transport.mailets.BayesianAnalysisFeeder
>
> Only 4 classes, and they are bound to
> excalibur/cornerstone/avalon-framework...
>
> If you're willing to refactor them to remove the avalon/excalibur
> dependencies then maybe they can live alone, otherwise I hardly see any
> advantage in moving it out.

IIRC the avalon code is used for data access. it would be easy to
replace this with JPA (say) or JCR.

- robert

---------------------------------------------------------------------
To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
For additional commands, e-mail: server-dev-help@james.apache.org


Re: [mailet] Creaet Bayesian-Mailet product? [WAS Re: [jira] Commented: (JAMES-621) Move BayesianAnalyzer to extra project]

Posted by Stefano Bagnara <ap...@bago.org>.
Robert Burrell Donkin ha scritto:
> opinions?

Unless there is someone willing to work on that code I think it is bad
to move it out.
Otherwise we'll end up with a microlibrary for the ClamAVScanner, for
the SpamAssassinInvoker, for the NetworkMatcher and so on.

You proved that you take care of releasing what you proposed to extract,
so I believe in you if you want to follow this path.

We have
org.apache.james.util.BayesianAnalyzer
org.apache.james.util.JDBCBayesianAnalyzer
org.apache.james.transport.mailets.BayesianAnalysis
org.apache.james.transport.mailets.BayesianAnalysisFeeder

Only 4 classes, and they are bound to
excalibur/cornerstone/avalon-framework...

If you're willing to refactor them to remove the avalon/excalibur
dependencies then maybe they can live alone, otherwise I hardly see any
advantage in moving it out.

Stefano

> - robert
> 
> On Tue, Nov 18, 2008 at 5:59 PM, Robert Burrell Donkin (JIRA)
> <se...@james.apache.org> wrote:
>>    [ https://issues.apache.org/jira/browse/JAMES-621?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12648676#action_12648676 ]
>>
>> Robert Burrell Donkin commented on JAMES-621:
>> ---------------------------------------------
>>
>> We could break out the analzers and mailets out into a separate product under mailets without too much difficulty.
>>
>> We would probably need to fork SQLUtils (but it's a simple class).
>>
>>> Move BayesianAnalyzer to extra project
>>> --------------------------------------
>>>
>>>                 Key: JAMES-621
>>>                 URL: https://issues.apache.org/jira/browse/JAMES-621
>>>             Project: JAMES Server
>>>          Issue Type: Wish
>>>            Reporter: Norman Maurer
>>>            Priority: Minor
>>>
>>> It whould be cool to move the BayesianAnalyzer stuff to an extra subproject. This whould make it easy for other developers to use our code.
>> --
>> This message is automatically generated by JIRA.
>> -
>> You can reply to this email to add a comment to the issue online.
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
>> For additional commands, e-mail: server-dev-help@james.apache.org
>>
>>
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
> For additional commands, e-mail: server-dev-help@james.apache.org
> 
> 


---------------------------------------------------------------------
To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
For additional commands, e-mail: server-dev-help@james.apache.org