I’ve been the lead developer on a desktop project for over a year now, and now we’re about to take it to the web. No problems there, and I’ve frequently asked for advice about this here. One of the web guys (who can’t code worth anything; he has volumes of un-indented code, which in my book is inexcusable) has sort of taken the project over, which is OK by me as he’s done the work of talking to the client and he certainly knows more about Oracle databases than I do.

However, I am still going to be the primary (perhaps only) coder on this project, so once we get development started it’s going to pretty much be my game. The solution I’ve been intending to use (well known to the web guy) is jsps for the UI and PL/SQL for the server side code. Then today he comes in and tells me that he and my other boss (my “real” boss if you will) have decided between them that “we” aren’t going to use jsps and instead are going to do Oracle portlets, which I honestly know nothing about. The reason for this being that there is little jsp experience on our development team (there are about a half-dozen of us who work on various projects) and he and the other web guy have done tons of portlets (which the other guy has complained to me about in great detail).

Perhaps there is some good solid logic here; I wouldn’t know, not knowing anything about Oracle portlets. However, the thing that really bugs me is that I wasn’t consulted AT ALL. In fact he KNEW I’d rather be working in jsps (because I’d be working in Java, much closer to my favored Delphi) but decided to progress without me. My “real” boss knows nothing about anything web (he’s a tried and true Oracle database programmer who only does forms), so basically whatever the web guy said is what we’re going to do.

Now I’m furious because a) I’ve already developed jsp screens that we used in our design specification document for the client, and these could be easily fleshed out to work with real data, and b) I feel (since I’m the primary coder, I was in charge of the desktop version, and I know more than anyone else here about how this is all supposed to work) that I should have been in on any discussions regarding development path, even if my ideas weren’t used. As it stands now, I don’t know what the best solution is because nobody gave me the opportunity to find out. And as our “web” guys experience is completely with Oracle Portal, I’m not highly confident that he really has a firm grasp on the best solution anyway. In fact, he told me that jsps are probably superior, but that “for the team” the best solution is to use portlets (since I’m the only one who knows anything about jsps). I asked if there could be further discussion and he said “no, this is how it’s going to be”.

This is the most angry I’ve been in a while. I’m willing learn new stuff, and I’ll grudgingly even implement something I don’t think is a good solution, but having my opinion completely disregarded in favor of someone’s who write shoddy code is too much to over come. Am I overreacting? Is this a case of my arrogance and familiarity with a solution clouding my judgment, or do I have legitimate concerns?

I know my opinion doesn't mean much, but here's what I'd do.

Choice #1: I'd tell the shoddy-code guy, "well fuck it. You're the Oracle Portlet expert, let's see what you've got in mind." And try to shift the "primary coding job" over to him. This sounds like a power struggle, and he's just trying to get it done his way. But if he's forced to do the primary coding because he's the one with all the experience with the portlets, then he won't like that, because it'll make him feel like the bitch. While he does all the hard work, you standby and watch.

Choice #2: if choice #1 isn't feasible, or you really want to do the primary coding yourself, then talk directly to your real boss. Tell him how you think it should be done. I'm sure he would understand the situation. If you're the one doing the majority of the work, then it should be up to you how it is done, not up to some dumbass shoddy-code guy.

Choice #3: Buy a semi-auto rifle from a pawn shop and shoot the place up. (kidding)

Forgive me for not knowing a thing about jsps, so this may be a dumb reply... Is training the rest of the developers an option? I think the decision should be (at least partly) based on the future development of the project. Is there a greater benefit by using jsps vs Oracle portlets, and is it financially and time wise feasible to get others trained? Sounds like someone is dropping the ball somewhere. That sucks...

No, it’s not a dumb reply. As a matter of fact, one of the other developers is also learning more about jsps (the second of our two “web guys”). The “web guy” in question, however, refuses to have anything to do with them.

To be honest I don’t know how much of an advantage jsps offer over portlets because I’ve never done any portles. But neither has he done jsps. The thing is, with such a void of knowledge, how can we hope to make a good design decision if we don’t at least include all viewpoints in a decision?

Roughly put, though jsps are faster, will be quicker to develop (for me, anyway), and offer more flexibility. The benefit of portlets are that more of the team is familiar with them (only 2 people, actually, so it’s not THAT big of an advantage) and apparently security is built in. Truly I’m not totally certain how to implement solid security from within jsps, so that might be a killer benefit. But again, how can we know that without any meaningful discussion?

While I understand your point exactly, it seems to me to be that if you leave, then there is nobody that can modify you code if the need arises. Besides, it's a lot easier to train one guy than it is 15.

For the record, it doesn't seem to me that any animosity is pointed toward you, it's just that the decision has more to do with what's right for the company, not what's right for the application.

All that being said, I'd be pissed too if that kind of decision were made without even asking me. But then, I'm the boss around here, so...

#2 is probably the choice I'll take, after I calm down a bit!

And I was wondering when you'd get around to option #3

As a slightly less lethal alternative to #3, If I knew you didn't have moral objections to it, I'd suggest you download a load of porn onto his computer and idly comment to the boss that you wish he's quit looking at it during work hours.

it seems to me to be that if you leave, then there is nobody that can modify you code if the need arises

I think that any even vaguely competent programmer should be able to pick up the rudiments of a new language in no time at all. I don't think lack of familiarity with a language should be any stumbling block at all.

However, if the basic tenets of JSPs are wildly different than those of portlets then that might be an issue, but I can't see how it would really be.

It also seems that the other developers' aversion to portlets makes their knowledge of it much less significant.

For the record, it doesn't seem to me that any animosity is pointed toward you, it's just that the decision has more to do with what's right for the company, not what's right for the application.

You're right about this, or at least the part about there not being any animosity toward me. A large part of my anger probably comes from my feeling (and I'm not alone) that this particular individual is incompetent, and therefore it is difficult for me to trust his judgment.

In EVERY situation that we've run into roadblocks during development, he always comes up with some "scheme" using technology with which he is familiar (you know: "how about we copy this into a file and perform XYZ on it, do a screen dump, parse the information with a batch file, and load it back into our application?") rather than listening to other people's suggestions who may have skills that he doesn’t possess. Normally I just ignore his input (a stand which has served me well), but now that he is moving more directly above me in this project I can't really do that. So perhaps the problem is more one of the fact that I don't trust him to evaluate our options fairly, and because of that lack of trust on my part I am very frustrated at not been given a say in how this is going to proceed. Either way I'm going to talk to my "real" boss about the situation in general. I do not want this person above me, and if he is officially going to be the lead on this project then I'm going to calmly and rationally make my feelings known about his ability to lead.

I would say it sounds like your real boss needs to more clearly define job roles, but it sounds like you should talk to your boss and tell him that you will likely be the primary developer and that you'd like to discuss the non-portlet option more. I think it's important to not make it look like you're just trying to overrule the other guy -- rather you want to open the discussion again. That will show them that you actually care about the decision.

[oracle-rant]
As far as using Oracle stuff vs JSP -- I haven't used portlets, all I can say is that Oracle may have the fastest database in the world, but they suck at writing applications for developers. I get so angry every time I work with Oracle and get a -234091 error with a cryptic error message (i mean really, what year is this? you can't tell me the name of the column that had a problem?). With JSP/Servlet/J2EE, you have many options of deployment platforms, you have many open source options, you have a ton of community tools and IDE integration tools for development, and you don't get vendor lockin like with everything Oracle touches. If you ever decided that perhaps you didn't need or want to be paying $30k a year in Oracle licensing, if you've probably written your application with j2ee, you can switch out your database. If you ever need to deploy your app on someone else's system, you can cut the deployment cost WAY down (you can deploy on Apache + Tomcat + PostgresSQL, or maybe they want to use SQLServer).

In general, the more standard something is, the better, in my mind. If it says Oracle in the title, you can probably count on it screwing you later on.
[/oracle-rant]

That being said, we use Oracle here for our primary database (like I said, it's fast), but we also deploy on pgsql as on option.

ms

I think it's important to not make it look like you're just trying to overrule the other guy -- rather you want to open the discussion again. That will show them that you actually care about the decision.

Hmmm, that's probably good advice, rather than questiong this other guy's competency. I think I'd be OK with whatever we decide as a group, as long as I'm involved in the discussion.

all I can say is that Oracle may have the fastest database in the world, but they suck at writing applications for developers.

Amen to that. I have said this very thing over the last few weeks as I've used some of their design tools. I can't believe how poorly their software is implemented.

If it says Oracle in the title, you can probably count on it screwing you later on.

True, but not an argument I can use with my boss!

I'd have to agree with mschrag on this one. JSP with Javabeans can be a totally independant layer on whatever database you choose, while Oracle portlets will probably be tied just to Oracle databases. Also compare the number of Java developers with Oracle specific developers and let them know if you were to leave there would be tons of people available to modify your code. Also, if you ever wanted to sell the technology or farm it out to a contracting agency it would be easier to do in Java than portlets.

I'm sure if you found a way to prove your case in a monetary way it would also help. That's what I do at my workplace and it makes a powerful point. "If we buy this Rational software, it will cost us XXXXXX dollars, while if we were to use CVS and Bugzilla it costs us nothing"

Well if you go to #3 I know a guy that can hook you up cheap no questions asked, PM me to get the details haha

Here’s a follow up in case anyone’s interested: It looks like someone’s about to have his wings clipped. My “real” boss told me that this decision had not been set in stone, and that he’d wanted myself and the other guy to discuss it more. Clearly that did not happen. The other guy is the lead of this project officially, and I do have to go with what he says. However, this really is only meant for project direction, to keep it on track and all. Since I am the primary developer I should be able to make most of the development decisions. He told me he’d clarify this with the other individual, and that if I have any problems to bring any issues to him directly. If the other guy takes a hard stance on anything, I do have to do what he says, but at least there’s recourse if there is a really bad decision being made. So we may end up using PL/SQL to do our web development after all, but at least I’m going to have my say. And now I know I can go to my boss with development issues. Looks like I won’t have to resort to option #3!

For what it's worth, you might consider putting together equivalent "Hello, world" apps in both the J2EE and Oracle frameworks. Then, you could start coming up with quantitative metrics about how long it would take you to finish the project. Another good quantitative number might be the amount of activity you can find on relevant chat boards (# of active users? posts per day?) where you might go asking for help or even the number of relevent books in the bookstore. Such "hard" numbers will lend more weight to anything qualitative you might have to say (such as commentary on the quality of 3rd party tools or your own personal preferences).

Find another job, quit, and do a head job on the f****** 'cos he's a cunt who wants to flesh out his CV for the next job.

Retain your integrity. Use bad language.

steady on Rob.