Call for Java programmers
John Metta
This is ridiculous. I've been spinning my wheels way too much on the
codebase because, as it turns out, it's been far too long since I've
done any serious Java programming. I thought I'd be able to jump back in
cleanly as with PHP and other languages, but clearly that's not the
case. I really like what BigBark has to offer, but it's proving too much
for me at this point and few others have jumped into the code.
So, A.J. and I have discussed what the best steps are and decided to
bring it to the group. It seems that without a strong Java programmer
with the time to take the reigns of the coding, we might be in a tight
management situation even after the BigBark codebase is running.
So the call is for a java programmer who can dedicate the time to
quickly bring the BigBark code up to speed and can also help maintain
the code in the short term until others come up to speed. Any takers?
Failing that, A.J. and I mentioned finally following Jeff Martens'
advice and stepping back to use Kyle's or another option to work with
the database that A.J. has. Depending on the language, Kyle's code might
be more manageable without strong Java coders on the team, and more
extensible with multiple developers.
I've held tightly to the idea of BigBark because it was supposed to be
nearly ready to go, but have to admit defeat before October breezes by.
Sound off now if you have thoughts on this either way.
-J
Patrick Lightbody
I saw you posted to the PJUG list. I got one email from someone I know
from the Struts world (didn't even know he was local!). Have you
received anymore responses? You might want to send a followup email
with a link to the getting started wiki page I created last week, just
to lower the barrier to entry a bit more.
Also - I want to make clear: I'm happy to be involved in hands on
coding, but I just can't *lead* the effort through to completion. But
in terms of getting others up to speed, mentoring on the codebase,
etc... I'd be happy to do that.
Patrick
Craig McClanahan
On Oct 15, 4:51 am, "Patrick Lightbody" <pligh...@gmail.com> wrote:
> John,
> I saw you posted to the PJUG list. I got one email from someone I know
> from the Struts world (didn't even know he was local!).
> Have you
> received anymore responses? You might want to send a followup email
> with a link to the getting started wiki page I created last week, just
> to lower the barrier to entry a bit more.
>
> Also - I want to make clear: I'm happy to be involved in hands on
> coding, but I just can't *lead* the effort through to completion. But
> in terms of getting others up to speed, mentoring on the codebase,
> etc... I'd be happy to do that.
>
others ... its just that (per my longer email earlier) I would like to
see it based on an assumption of wild success, rather than "just
another webapp that will not scale if it *does* become popular." I'm
also a getting a bit allergic to forty gazillion framework layers (and
a 9mb war file) for something that, right now, isn't all that
complex. Finally, there are leverage opportunities here if we look
(probably later on, based on availability of coding effort)
integration with other social networking sites.
Any comments on my suggested approach?
Craig
> Patrick
Patrick Lightbody
I'm definitely flexible - really, it's up to whoever steps up. My only
advice there would just be to be careful: even simple projects end up
being large than you originally expect and we've got a fairly complete
package, so I would just watch out for a case of throwing the baby out
with the bathwater.
That said: can you give a bit more detail about what your approach is?
I hear your concerns, but I don't know what you'd like to do about
them. The stack selected is, for the most part, actually fairly
"light" in terms of jars/bytes/etc:
* Guice is three dependencies: guice.jar (500K), aopalliance.jar
(5K), and guice-servlet.jar (15K)
* Stripes is two jars: stripes.jar (500K) and commons-logging.jar (59K)
* Warp-Persist (the Hibernate/Guice integration) is a single jar at 66K
* Hibernate is a touch "heavy", with several jars (hib core, hib
validations, c3p0, cglib, asm, etc) all weighing in ~3MB or so
Can you help me better understand what you're concerned about? Disk
space isn't something that bothers me. Complex dependencies do bug me
a bit more, but in terms of off the shelf options, the ones selected
here are fairly low on deps (minus Hibernate).
Is your desire to code the thing up using Serlvets, JSPs, and straight
JDBC because you find it to be a simple app?
You also mentioned "scale". Are you talking about performance or the
ability for the code to scale up with features?
Patrick
John Metta
enthusiasm come into the effort. First off, I'd like to move this
discussion to the ORBlogs list to reduce duplicated traffic.
A few more things.
Craig, I agree on your assessment, and am super excited because that is
the path that ORBlogs wants to go. However, we want to get the BigBark
code up and fully functional so that we have a good site running and
generating traffic WHILE we are working on such an effort. Reasons for
this are as follows.
0) BigBark is basically ready to run. We need to dust it off and turn
the key, this makes it an optimal instant, if temporary, solution.
1) There are at least 1800 blogs currently indexed which are now not
served. Many of these folks are, understandably chomping at the bit
since we (i.e. I) originally shot for Oct. 1st.
2) There are a large number of blogs (I have 4 alone) that are not
indexed which need to be added to ORBLogs ASAP.
e) Having BigBark up- and experiencing the slam of hundreds of bloggers
hitting it with full expectation- will not only give us juice, it will
give us data and information on what we want and need in the larger effort.
3) The thought was that it was a mere few days to get BigBark running
with strong Java people, so it would not take away from the longer term
effort.
4) Once BigBark is running, we can spend the actual time to design an
excellent system while blogs are being indexed and traffic is being
generated. this will make the blogging community very happy.
Patrick can discuss the timeline for BigBark better than I, but it
really seems as though it will be trivial to get it up. This is
important, I can't stress enough how many people are waiting for this to
come back online. Once this is up, I've got some ideas to add to Craig's
wonderful list. I'm excited, this is going to be great.
Cheers,
-J
Lyle T. Harris
The getting started page should include the Maven option
to run it of "mvn jetty:run-war" which is much handier.
Per Bill's request I'd like to help out where/when I can.
What specific areas need addressing? The bigbark.net site
seems to be currently up and running.
Operators are standing by... ;-)
Regards,
Bill Jackson
I also finally got it running on my system yesterday. I also created
and resolved an issue and checked in a few trivial changes before I
ran out of time. :)
Anyway, Patrick had a list of things to be done to make the BigBark
code ready for ORBlogs.org. I think you can find it in this mailing
list's archive. I'll see whether I can find it when I get back to my
computer (writing this on my phone) and save it to the wiki, then we
can break it out into individual issue/enhancement records.
Since I didn't have time to start configuring my instance, the only
thing I can think of right now is the reskinning needed. :)
-Bill
> begin:vcard fn:Lyle T. Harris n:Harris;Lyle T. org:Sun Microsystems
> Inc.;Software Integration Engineering adr;dom:;;;Portland;OR
> email;internet:lyle....@sun.com title:Principal Engineer x-
> mozilla-html:TRUE version:2.1 end:vcard
Patrick Lightbody
There are likely a lot of bugs in the code that cropped up when the
stack was modernized to use Guice/Stripes/Hibernate. So I'd say the
best way to get started is to do one of three things:
1) Look around for the bugs and close them as you find them.
2) Work on getting the UI updated so it says "Oregon Blogs" rather
than "political blogs" :)
3) Work with AJ and John to see about mapping in the database dump
they got from the original site and ensuring that all the feeds can be
parsed in this codebase.
I'd say what we should do is split up these three top-level tasks and
assign an owner to each one. Then they can likely be chased down in
parallel.
Patrick
Patrick Lightbody
What's your Google Code username? I'd be happy to get you SVN access.
Patrick
John Metta
to work on #3. I'm also working on some non-programming issues (of
course), but want to keep the bottlenecks to a minimum. I'll make sure I
have the latest DB version from AJ, and if someone wants to take the
Java side of DB mapping, let me know.
-J
Patrick Lightbody
How's the DB mapping stuff going? Which approach do you think you'll
pursue? Mold AJ's schema to fit the BigBark schema, or modify the
BigBark code to work with AJ's schema?
My gut says that option #1 is probably easier, but I'll defer to those
who do the actual work of course :)
Patrick
A.J.
know what you need and I will make it happen.
A.J.
John Metta
easiest. We can re-evaluate once things are running and we better know
our needs.
I meant to get with Bill earlier to ask some questions (duh, actually
didn't think to ask Patrick), but didn't get to it and I'm out this
weekend with family. AJ, I was assuming you'd be booked working on your
new OS so was taking the DB work. Shouldn't take too much time. Patrick,
did you actually have a formal database schema in the code? Sample.sql
is all I saw:
http://orblogs.googlecode.com/svn/trunk/src/main/resources/sql/sample.sql
Cheers,
-J
Patrick Lightbody
http://orblogs.googlecode.com/svn/trunk/src/main/resources/sql/schema/postgres/schema.sql
But it's actually not totally right, since it's the postgres schema
and when I updated the code to the latest tech, we switched over to
MySQL. As it stands now, Hibernate is auto-creating the schema, so
we'd have to do a dump of the schema right after Hibernate creates it
to get the formal DDL.
However, that link is probably a good starting point to get a high
level picture of the schema. Just don't trust it for accuracy under
MySQL.
Patrick