Accessing pre-existing data

View: New views
6 Messages — Rating Filter:   Alert me  

Accessing pre-existing data

by Daniel J. Lauk-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hi list.

The next problem I have to face in evaluating the frameworks, is: "How
do I access pre-existing data?"
In my special case, we have a centralized data store for "personnel
master data", that I need to access. This resides on a database server
that won't necessarily be the one that the app is running on.

I am sure, I can figure out how to fine tune hibernate to manage two
connections, but how much does this impact the remaining Trails stack?

Another application area would be, to use Trails to re-write existing
web-applications (e.g. because the code is *so* old or messed up).
The only approach that comes to my mind (I am not into java frameworks
yet -- just starting) is, create the new business logic from scratch
with trails and then (backup and) migrate the existing data (similar
to the model evolution problem discussed before).

Kind regards,
Daniel

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email



Re: Accessing pre-existing data

by Alejandro Scandroli :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hi Daniel

Pre-existing data is not an issue.
On the other hand working with two data sources can be tricky.
If you can trick hibernate to manage two or more different db
connections at hibernate level Trails won't notice it, and it will
work without problems.
If that's not possible and you only need the second connection to
validate your users, you can create your own UserDetailsService to
serve your user credentials.
If that's not enough you can use two persistence services but that
means that you will have to duplicate quite a lot of configuration,
and I haven't tried it with Trails 1.2
Please if you are just starting with java web frameworks, take a look
at: http://www.trailsframework.org/Read+me+first


Regards.
Alejandro.

--
Alejandro Scandroli - http://weblog.amneris.es/
Amneris: We build process-driven web applications.
http://www.amneris.es



On Thu, Apr 17, 2008 at 11:31 AM, Daniel Lauk <daniel.lauk@...> wrote:

> Hi list.
>
>  The next problem I have to face in evaluating the frameworks, is: "How
>  do I access pre-existing data?"
>  In my special case, we have a centralized data store for "personnel
>  master data", that I need to access. This resides on a database server
>  that won't necessarily be the one that the app is running on.
>
>  I am sure, I can figure out how to fine tune hibernate to manage two
>  connections, but how much does this impact the remaining Trails stack?
>
>  Another application area would be, to use Trails to re-write existing
>  web-applications (e.g. because the code is *so* old or messed up).
>  The only approach that comes to my mind (I am not into java frameworks
>  yet -- just starting) is, create the new business logic from scratch
>  with trails and then (backup and) migrate the existing data (similar
>  to the model evolution problem discussed before).
>
>  Kind regards,
>  Daniel
>
>  ---------------------------------------------------------------------
>  To unsubscribe from this list, please visit:
>
>     http://xircles.codehaus.org/manage_email
>
>
>

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email



Re: Accessing pre-existing data

by Daniel J. Lauk-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hello Alejandro.

Thank you once again for your response.

> Pre-existing data is not an issue.

Good. One worry less on my list :-)

> On the other hand working with two data sources can be tricky.
> If you can trick hibernate to manage two or more different db
> connections at hibernate level Trails won't notice it, and it will
> work without problems.

A colleague of mine is quite good at working with hibernate. I'll let
you know, if I find a solution to that problem.

> If that's not possible and you only need the second connection to
> validate your users, you can create your own UserDetailsService to
> serve your user credentials.

It's not used for authentication. It's used to serve "personnel master
data", e.g. who is in which department, which division does the
department belong to, accounting information, report-to-chains etc.

For authentication I'd use other tools (i.e. we run a CAS server and
could also use shibboleth).

> If that's not enough you can use two persistence services but that
> means that you will have to duplicate quite a lot of configuration,
> and I haven't tried it with Trails 1.2

OK. Good to know, that -- just in case -- there are options left to try.

> Please if you are just starting with java web frameworks, take a look
> at: http://www.trailsframework.org/Read+me+first

I did read it first. And then the quick start, the overview, most of the
user's guide and the articles. I read most of the information on the web
site.

Currently my task is evaluating -- in a sense of "read specs" and/or
"try a tutorial". I got my feet wet, but my colleagues ask questions,
that I could not answer from the documentation. That's why I pass those
questions on to the mailing list. I am sorry, if you consider my
questions inappropriate.

Basically, all I have to do, is find out, what the different candidates
(amongst them Trails) are capable of and how well they support different
tasks. Then I'll present that to the "evaluation board" and they'll pick
something. And only *after that* I'll go through the trouble of learning
the necessary tools (maven2, ant, hibernate, tapestry, ...) in detail.

As far as I can tell, I suppose, that once the decision is made, I'd
even be allowed to contribute to Trails' documentation during work time
instead of writing "internal" docs. My boss is quite open minded and
generally likes open source and supports giving back to the community.
He is also a Java enthusiast, so I think that Trails has a good chance
of being chosen.

Many thanks and kind regards,
Daniel

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email



Re: Accessing pre-existing data

by Alejandro Scandroli :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hi Daniel

> I am sorry, if you consider my questions inappropriate.

No, please, you have have been very polite and concise. I really
appreciate that. If you thought I was considering them inappropriate I
have to blame my lack of good English skills. I'm sorry.

Finally (to be completely honest)  one more thing to consider in your
evaluation.
The soon to be released 1.2 version of Trails is based in Tapestry 4.
There is a new Tapestry version, Tapestry 5, which is beautiful and
very fun to code with, but sadly It's not fully backward compatible
with Tapestry 4. That means that the next version of Trails (2.0?)
will have to be rewritten almost from scratch. IMO that's not a bad
thing at all, but that's something you need to know to consider in
your evaluation.

Trails certainly needs some help in the documentation department so
I'm looking forward to your contributions.


Saludos.
Alejandro.

--
Alejandro Scandroli - http://weblog.amneris.es/
Amneris: We build process-driven web applications.
http://www.amneris.es







On Fri, Apr 18, 2008 at 12:12 AM, Daniel J. Lauk <daniel.lauk@...> wrote:

> Hello Alejandro.
>
>  Thank you once again for your response.
>
>
>
> > Pre-existing data is not an issue.
> >
>
>  Good. One worry less on my list :-)
>
>
>
> > On the other hand working with two data sources can be tricky.
> > If you can trick hibernate to manage two or more different db
> > connections at hibernate level Trails won't notice it, and it will
> > work without problems.
> >
>
>  A colleague of mine is quite good at working with hibernate. I'll let you
> know, if I find a solution to that problem.
>
>
>
> > If that's not possible and you only need the second connection to
> > validate your users, you can create your own UserDetailsService to
> > serve your user credentials.
> >
>
>  It's not used for authentication. It's used to serve "personnel master
> data", e.g. who is in which department, which division does the department
> belong to, accounting information, report-to-chains etc.
>
>  For authentication I'd use other tools (i.e. we run a CAS server and could
> also use shibboleth).
>
>
>
> > If that's not enough you can use two persistence services but that
> > means that you will have to duplicate quite a lot of configuration,
> > and I haven't tried it with Trails 1.2
> >
>
>  OK. Good to know, that -- just in case -- there are options left to try.
>
>
>
> > Please if you are just starting with java web frameworks, take a look
> > at: http://www.trailsframework.org/Read+me+first
> >
>
>  I did read it first. And then the quick start, the overview, most of the
> user's guide and the articles. I read most of the information on the web
> site.
>
>  Currently my task is evaluating -- in a sense of "read specs" and/or "try a
> tutorial". I got my feet wet, but my colleagues ask questions, that I could
> not answer from the documentation. That's why I pass those questions on to
> the mailing list. I am sorry, if you consider my questions inappropriate.
>
>  Basically, all I have to do, is find out, what the different candidates
> (amongst them Trails) are capable of and how well they support different
> tasks. Then I'll present that to the "evaluation board" and they'll pick
> something. And only *after that* I'll go through the trouble of learning the
> necessary tools (maven2, ant, hibernate, tapestry, ...) in detail.
>
>  As far as I can tell, I suppose, that once the decision is made, I'd even
> be allowed to contribute to Trails' documentation during work time instead
> of writing "internal" docs. My boss is quite open minded and generally likes
> open source and supports giving back to the community.
>  He is also a Java enthusiast, so I think that Trails has a good chance of
> being chosen.
>
>  Many thanks and kind regards,
>
>
>  Daniel
>
>  ---------------------------------------------------------------------
>  To unsubscribe from this list, please visit:
>
>    http://xircles.codehaus.org/manage_email
>
>
>

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email



Re: Accessing pre-existing data

by Daniel J. Lauk-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hey Alejandro.

2008/4/18, Alejandro Scandroli <alejandroscandroli@...>:
> Hi Daniel
>
>
>  > I am sorry, if you consider my questions inappropriate.
>
>
> No, please, you have have been very polite and concise. I really
>  appreciate that. If you thought I was considering them inappropriate I
>  have to blame my lack of good English skills. I'm sorry.

Please don't be sorry. There's absolutely no need. I wasn't sure, what
you meant, because I'm no native speaker as well, so the "error" just
summed up, I guess. :-)
Anyway: Now we're clear on that. Won't be a topic again.

>  Finally (to be completely honest)  one more thing to consider in your
>  evaluation.
>  The soon to be released 1.2 version of Trails is based in Tapestry 4.
>  There is a new Tapestry version, Tapestry 5, which is beautiful and
>  very fun to code with, but sadly It's not fully backward compatible
>  with Tapestry 4. That means that the next version of Trails (2.0?)
>  will have to be rewritten almost from scratch. IMO that's not a bad
>  thing at all, but that's something you need to know to consider in
>  your evaluation.

Thanks for that insider's view. I'll add it to my summary and thoughts.
We have a strong Tapestry antagonist in the department but also a
Tapestry protagonist (the hibernate guy).

>  Trails certainly needs some help in the documentation department so
>  I'm looking forward to your contributions.

I hope I didn't promise too much. I do not know, if Trails will be the
"winner", and if not I certainly won't have an opportunity to
contribute.
In private life I stick to script languages because the cheap web
hosters offer them for free ;-)

Best wishes,
Daniel

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email



Re: Accessing pre-existing data

by Daniel J. Lauk-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hi.

>  On the other hand working with two data sources can be tricky.
>  If you can trick hibernate to manage two or more different db
>  connections at hibernate level Trails won't notice it, and it will
>  work without problems.

I talked to our hibernate guy. He said that it is possible to
configure hibernate to work with different db schemes. He's using such
a configuration in one project.

Kind regards,
DJ

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email