[Jpolls-developers] Pipeline

[Jorge L. de C. <jo...@ho...>]

Thanks for your suggestions and ideas Osvaldo,

I've added your suggestions to the "suggestions" list on the @todo file. =
Perhaps I should have told you lot what I had in the pipeline as well.
My short/medium term add-ons are:

1) Servlets accepting id and outputting poll HTML with CSS support. The =
servlet responds to any kind of device that communicates using the right =
interface. As of yet, you enter the main URL to access the poll's page. =
The purpose of the servlet is to enable automated systems to access =
polls as well or users to be able to insert code similar to <a =
href=3D"http://www.jpolls.net/getPoll?id=3Dwhatever"> to have their poll =
data streamed into their page section of choice.

2) Integrate user management. I have got an user management module from =
another project on another company that I will *change* and donate to =
the SF struts group. Integration of this module should be =
straightforward. The purpose is to keep anonymous polls but also to =
allow people that signin to be the only editors of the polls they =
create. Perhaps user hierarchies could also be accounted for.

3) Proper statistics module, which would require a few more persistence =
services. Also, some of the methods on the persistence services =
interface need rewriting/refactoring/split.=20

4) Currently, underlying support for ANY multimedia resource is =
available for the poll options. This means users can add images, audio, =
or video to their poll options/questions, as of now. I have disabled the =
video and audio because we need a small self-contained player for those =
file types (you don't really want the user to open MediaPlayer or =
whatever to see the multimedia resource...). Just when I was finishing =
my small applet audio player, a friend pointed me to an existing one =
-tested, working, smaller. Damn. It looks good but now I've been =
thinking a better looking one can be achieved with Flash. So I'm putting =
myself into that task as I write this :-) The next version will have an =
audio player after the poll option whenever an audio file was uploaded =
for it. But the version after the next will hopefully have a niiiice =
looking small Flash player for audio AND video!

5) Offer a few more of the easy goodies such as overlaid charts, charts =
for this, charts for that, you get the point. One of the things I like =
about the direction this is taking is the inherent and potential visual =
appeal.

6) Add a configuration editor and module. Currently, I've got a XML file =
with some properties that I edit to have some global properties change =
effects. I'd rather do this on an HTML form editor with dropdown menus =
and checkboxes.

7) (Piotr's) Add a J2ME module that integrates smoothly and allows J2ME =
devices to be part of an enormous pool of polls! In other words, get =
this component mobile (mobile IS trendy these days).

Long term plans:
1) Currently, although support for any number of poll options is =
possible and advertised, this infrastructure is essentially a static =
one. In the future we need a dynamic structure, i.e., separate poll =
options from the poll entity and create a 1-2-Many relationship between =
polls and poll options. This saves space (only those answers inserted =
and available are kept), and gives the user better flexibility regarding =
the number of options they want to provide per poll. It also seems to me =
it is an unique feature. I haven't seen any already made component with =
such flexibility.

Short/Long term constraints:
1) Keep the core small. Currently, it has ~140KB. I think a fully =
featured component can be achieved within the ~250KB-300KB size range.
2) Look for bottlenecks, concurrency access and violations, code =
optimizations, especially in the persistence layer. The next layer =
important to optimize is the charts package. I _doubt_ there is any =
optimization needed for the actions package (!) -code there has been =
thoroughly tested in other packages in other modules in other projects. =
It is stable as a rock, elsewhere.

Although the above items are numbered they are not to be in any peculiar =
order. I would very much appreciate hearing your ideas! So if you have =
anything to add on or change, let us know.
                                                                         =
                 ...
My comments to Osvaldo's suggestions are:
-Docs YES! It has been a constant white noise in the back of my head. =
Besides, OpenOffice these days allows me to write some proper doc and =
have it saved as .pdf. I know, I know, I've got no excuses...
-Regarding the Jax-RPC, by prevalence layer I guess you mean =
presentation layer :-). It does not make much sense as another =
persistence option. I thought about web-services when the word was =
trendy (is it still?). After doing some work with bloody SOAP on .NET =
and supervising work with bloody SOAP on J2EE, I can't say I am at all =
pleased with what can be achieved and how. It is an open option though, =
and I wouldn't say no to a free working module :-)
It would be on my not-for-this-year-options list, though.
-I admit security has not been of my initial concern. This started as a =
tutorial application to illustrate advanced Struts features, as I =
initially agreed with Ted Husted (one of Strut's creators). Free =
anonymous polls didn't require that much security in the beginning. =
However, thanks to them struts features and to the modular design (which =
still has plenty of room for improvement), security can be added easily. =
I think, however, that only a few places and a few selected =
transmissions would require such an effort.

Ok, prolix as usual, these are my 50cents.

Cheers
j.
      I=B4m from Brasil, my native language is portugese but i=B4ll try =
to speak english ;-).=20
      I think i=B4m the newer member of the team,  therefore,  i=B4d =
like to share some thoughts and questions about this project with every =
one of you:
      1 - i like Java, this is my "GANHA P=C3O".
      2 - i belive on this project.
      3 - i felt lack of project documentention.(i think it=B4s very =
important to the people know what does it do exactly)
      3 - What you think to use the Jax-Rpc on the prevalence layer? =20
      4 - What you think to use the crypt algorithms(java.security) to =
secure the data on the net?=20
      Well, this is some of my thoughts regarding the projetct. I=B4d =
like to know what you think about?=20
  Att,=20
      Osvaldo