> I'd like feedback on my talk, please. It'll help me improve my
> presentations. Specific areas that I'd like input on are technical
> depth of the talk, whether the examples were useful, would you like
> similar talks in the future, do you want more/less examples, what you
> think of the examples. Any other feedback would be welcome, of course.

> 2. The iBatis session which should have been called "object-data
> relationship modelling for ORM independent persistence" instead of
> iBatis.
> This is a good attempt to solve a problem in certain scenarios where
> on has an existing database structure and now needs to migrate to
> another persistence mechanism. The artifacts created by this tool can
> be used to generate wrappers for present and future persistence
> mechanisms. I was convinced by an example of a scenario where the
> speaker had first generated EJB code (about 800 odd tables), and then
> had to generate Hibernate code. Since both the code generators
> accepted different input, the input had to be recreated for the
> Hibernate code generator. This talk covered a tool that helps address
> this problem by helping one describe a data-object relationshiop in an
> ORM agnostic manner.

> In my opinion, this project can benefit a lot from looking at how
> Active Record works, and by having Active Record as one of the
> platform would use this ORM mapping.

> 3. The Digester session was different for me.
> Since I've not worked so much with CRUD/data intensive applications, I
> think of Digester as a tool to assemble together object hierarchies
> which have nothing to do with business domain models. I'd have
> definitely pitched Digester differently, with a different set of
> examples :).

> This was a rather long BOJUG meet. One the one hand it did go across
> some time, but on the other hand I learned during casual conversations
> that the discussions themselves were very engrossing. We finally had
> to stop the BOJUG user group meet, because the B.NET User Group folks
> wanted the room for _their_ meet :)

> Also, I forgot to get along the camera !

> -- Sriram

> On Sat, Sep 27, 2008 at 6:20 PM, Sriram Narayanan <sriram...@gmail.com>
> wrote:

>> This was a rather long BOJUG meet. One the one hand it did go across
>> some time, but on the other hand I learned during casual conversations
>> that the discussions themselves were very engrossing. We finally had
>> to stop the BOJUG user group meet, because the B.NET User Group folks
>> wanted the room for _their_ meet :)

> Shareable slides?

> Dear Sriram,

>  Your presentation was very good and informative. The examples which
> you took to explain load balancing, redundant systems were
> interesting. Your way of persentation and commmunication with audience
> was also good.

>  Raghavan's presentation on Jakarta Digestor was very good. It was
> more like a tutorial session.
> Thank you sir.

>  iBatis session provided insights to various issues in s/w
> development. The talk was comparitively difficult to follow and the
> presenter tried to cover a lot of topics. I will be nice if
> he can reduce code demos and give more importance to concepts.

> Thank you to all the presenters for finding time to share their
> knowledge.

> Vishal John
> Infineon Technologies, ITPL, Bangalore.
> Vishal.J...@infineon.com

> Sriram Narayanan wrote:
>> Some impressions:
>> 1. The load balancing session seemed to interest folks. I covered the following:
>> - binary and non-binary communication (http) between Apache and Tomcat.
>> - http load balancing between Apache and non-Java processes.
>> - Apache - Jetty communication over a binary protocol. (this can be
>> expanded to a load balanced communication as well)
>> - We saw how SSL termination can help separate SSL
>> encryption/decryption from the actual application servers and the data
>> processing.
>> - All of the above were on live production servers that I administer.
>> - We also covered what the terms load balancing, redundant systems,
>> hot standby, fail over, clustering, etc mean to various people.
>> - I'd wanted to cover hardware loadbalancing and failover as well, but
>> we didn't have enough time. Perhaps we could cover these some other
>> time.

>> I'd like feedback on my talk, please. It'll help me improve my
>> presentations. Specific areas that I'd like input on are technical
>> depth of the talk, whether the examples were useful, would you like
>> similar talks in the future, do you want more/less examples, what you
>> think of the examples. Any other feedback would be welcome, of course.

>> 2. The iBatis session which should have been called "object-data
>> relationship modelling for ORM independent persistence" instead of
>> iBatis.
>> This is a good attempt to solve a problem in certain scenarios where
>> on has an existing database structure and now needs to migrate to
>> another persistence mechanism. The artifacts created by this tool can
>> be used to generate wrappers for present and future persistence
>> mechanisms. I was convinced by an example of a scenario where the
>> speaker had first generated EJB code (about 800 odd tables), and then
>> had to generate Hibernate code. Since both the code generators
>> accepted different input, the input had to be recreated for the
>> Hibernate code generator. This talk covered a tool that helps address
>> this problem by helping one describe a data-object relationshiop in an
>> ORM agnostic manner.

>> In my opinion, this project can benefit a lot from looking at how
>> Active Record works, and by having Active Record as one of the
>> platform would use this ORM mapping.

>> 3. The Digester session was different for me.
>> Since I've not worked so much with CRUD/data intensive applications, I
>> think of Digester as a tool to assemble together object hierarchies
>> which have nothing to do with business domain models. I'd have
>> definitely pitched Digester differently, with a different set of
>> examples :).

>> This was a rather long BOJUG meet. One the one hand it did go across
>> some time, but on the other hand I learned during casual conversations
>> that the discussions themselves were very engrossing. We finally had
>> to stop the BOJUG user group meet, because the B.NET User Group folks
>> wanted the room for _their_ meet :)

>> Also, I forgot to get along the camera !

>> -- Sriram

>  Needless to say, yours was quite informative and I found it to be
> interesting. As I already mentioned (and Vishal had proved) the prerequisite
> of Apache HTTP Server was taken for granted from the audience. Though we
> can't cover the complete prerequisites, to certain extent it can be taken
> care.

>  Regarding Digester, What way I would have handled it? Thanks for the
> different dimension. It was my very first talk in BOJUG and i really dint
> have any idea on how it is expected from the people. Can you suggest me to
> take it up further? I would be glad to incorporate.

> Cheers,
> Raghavan alias Saravanan M.
> On Sat, Sep 27, 2008 at 6:20 PM, Sriram Narayanan <sriram...@gmail.com>
> wrote:

>> Some impressions:
>> 1. The load balancing session seemed to interest folks. I covered the
>> following:
>> - binary and non-binary communication (http) between Apache and Tomcat.
>> - http load balancing between Apache and non-Java processes.
>> - Apache - Jetty communication over a binary protocol. (this can be
>> expanded to a load balanced communication as well)
>> - We saw how SSL termination can help separate SSL
>> encryption/decryption from the actual application servers and the data
>> processing.
>> - All of the above were on live production servers that I administer.
>> - We also covered what the terms load balancing, redundant systems,
>> hot standby, fail over, clustering, etc mean to various people.
>> - I'd wanted to cover hardware loadbalancing and failover as well, but
>> we didn't have enough time. Perhaps we could cover these some other
>> time.

>> I'd like feedback on my talk, please. It'll help me improve my
>> presentations. Specific areas that I'd like input on are technical
>> depth of the talk, whether the examples were useful, would you like
>> similar talks in the future, do you want more/less examples, what you
>> think of the examples. Any other feedback would be welcome, of course.

>> 2. The iBatis session which should have been called "object-data
>> relationship modelling for ORM independent persistence" instead of
>> iBatis.
>> This is a good attempt to solve a problem in certain scenarios where
>> on has an existing database structure and now needs to migrate to
>> another persistence mechanism. The artifacts created by this tool can
>> be used to generate wrappers for present and future persistence
>> mechanisms. I was convinced by an example of a scenario where the
>> speaker had first generated EJB code (about 800 odd tables), and then
>> had to generate Hibernate code. Since both the code generators
>> accepted different input, the input had to be recreated for the
>> Hibernate code generator. This talk covered a tool that helps address
>> this problem by helping one describe a data-object relationshiop in an
>> ORM agnostic manner.

>> In my opinion, this project can benefit a lot from looking at how
>> Active Record works, and by having Active Record as one of the
>> platform would use this ORM mapping.

>> 3. The Digester session was different for me.
>> Since I've not worked so much with CRUD/data intensive applications, I
>> think of Digester as a tool to assemble together object hierarchies
>> which have nothing to do with business domain models. I'd have
>> definitely pitched Digester differently, with a different set of
>> examples :).

>> This was a rather long BOJUG meet. One the one hand it did go across
>> some time, but on the other hand I learned during casual conversations
>> that the discussions themselves were very engrossing. We finally had
>> to stop the BOJUG user group meet, because the B.NET User Group folks
>> wanted the room for _their_ meet :)

>> Also, I forgot to get along the camera !

>> -- Sriram

> On Tue, Sep 30, 2008 at 3:28 PM, Raghavan. 30May1981_GMail
> <raghavan.30may1...@gmail.com> wrote:
>> Dear Sriram,

>>  Needless to say, yours was quite informative and I found it to be
>> interesting. As I already mentioned (and Vishal had proved) the prerequisite
>> of Apache HTTP Server was taken for granted from the audience. Though we
>> can't cover the complete prerequisites, to certain extent it can be taken
>> care.

> Yes, I completely agree. In fact, I've replied as much a few minutes
> ago to another mail on this very thread.
> I also distinctly remember you telling me this after my talk.

> Here's how I used your feedback: The day before, I decided to bring a
> colleague up to speed with load balancing, and made it a point to run
> over the basics once just to ensure that we were both on the same
> page.

> I'll take care about this for all future presentations at the BOJUG.

>>  Regarding Digester, What way I would have handled it? Thanks for the
>> different dimension. It was my very first talk in BOJUG and i really dint
>> have any idea on how it is expected from the people. Can you suggest me to
>> take it up further? I would be glad to incorporate.

> The amount of content that you had for a first time BOJUG talk was
> absolutely fine.
> Here's what I liked:
> - examples every step of the way
> - examples for the direct API calls as well as the XML based decriptions.
> - all the examples worked, and there were no glitches at all at any point.

> Here's what was new to me:
> - Using the Digester as a data base population mechanism.
> I have always had colleagues who have taken an interest in databases
> and SQL queries, so I myself have focussed on everything else on the
> project.

> Here's how I'd have talked about Digester - only because of how I
> perceive Digester
> - An everyday case of accepting configuration directives from the user.
> - Enabling users to specify custom components
> - The need for wiring together all components at startup time rather
> than at programming time
> - The advantages of being able to do su (plug in different behavior,
> additional components, stripped down config files, sensible defaults)
> - Showcase how various well known open source projects use the
> Digester for the above
> - Tomcat's own self-assembly at startup-time, the defaults for web.xml
> (which we get away with not specifying), how objects' attributes can
> be set once more
> - How our editing of the Struts Config actually enables Struts to
> understand what it needs to do.
> - How expensive is the Digestion process actually ? (On my 2.2 Ghz P4
> with 2 GB RAM, the standard Tomcat startup with all the default
> webapps is approx 700 ms).
> - Using Digester ourselves (Here, I'd demonstrate my
> CustomClasspathClassloader and how I accept configuration for it from
> the user at startup time). Here's the code for those interested
> (https://issues.apache.org/bugzilla/show_bug.cgi?id=38223)
> - What Digester can and cannot do (demonstrate via the Tomcat admin application)
> - Digester vs OSGi (how the two cater to different audiences). This
> would be a brief slide only, since the OSGi, to me, is quite something
> else.

> So, there you have it. I'd have not thought about database centric
> applications of Jakarta Digester at all :)

> Something for you to think of for planning purposes:
> - If you recall, some of us in the audience asked you in between to
> show us code corresponding to what you'd just described. It's a good
> thing that you had the code, because we were able to follow you talk
> more closely. For future talks, please consider tailoring your slide
> order such that you can alt-tab into the IDE and show case code.
> Often, a code based equivalent of a slide can be very very
> illustrative.

>>  As we discussed, let us know when you planned to take up another
>> (rewinding) session on Tomcat Internals? I missed it last time and felt bad
>> for the same. It would be great if you can spare some time for the same.

> Ok, I should make some good slides for this too, and perhaps have more
> code examples.

>>  I shall be sending the slides and programs for Jakarta Digester to you as
>> you have suggested and I guess you can share it with others.

> Send it to Amit Saha too. I've still not figured out exactly where
> Amit hosts all the slides.