Day two of the uPortal Fundamentals Training I would deem a worthwhile success- I certainly got my money's worth. We focused mainly on XML and XSLT, and my pages of notes grew regarding how we could redesign the uPortal. It also dealt heavily with several problems in NextBrick's code, things I am not going to be happy to report. We've been for the past several months, effectivly hacking up a good portal, not effectivly transforming the content and information into something truly usable. There are effective ways to develop, and we took advantage of none of them. Administration is a simple thing, controlled ideally via properties and XML files... or at least it is supposed to be. Many of the personalizations we are paying for are built in to the portal, and even worse, many of these things don't need entire new channels. For example, our weblog did not need a custom channel. It could have gone through a simple XML/XSLT transformation. Furthermore, it would have been possible to render the blog's content entirely in the portal, which has always been our ideal goal. I'm kinda nervous about sitting down with my boss to discuss this laundry list of things, but there are so many things that would not only make our portal faster, but make it more reliable, allow for faster development, and reduce the HSDL's dependencies on external development solutions. I've avoided Java more out of a fear of spreading my langauge knowledge too thin than of anything else. I prefer to weild one or two languages well, but it looks like I'm about to pick up a few more. For those that have been keeping score since mechanicalruins,net, I like to mention my "comfort" zone on a sliding 1-5 scale, 5 being comfortably adept, 1 being ignorant. If it is not listed, I simply don't know it.
PHP (4. Up 1 point) (X)HTML (4. No change) XML/XSLT (2. Up 1 point) C++ (2. Down 1 point) Java (2. Up 1 point) PERL (2. No Change) Python (1. No change)
I've really stopped with c++. As much as I like it for applications, the amount of overhead required for true OOP functionality is cumbersome, ineffective, and largely impractical. I know that will probably draw some screams from people, but C++ was never designed to be an OOP language for web services, and I am to the bitter end a Web Developer. The more I learn about XML transformations though, the more interested I am becomming. I would absolutely love to see PHP implement an XSLT transformation and rendering engine. Nothing would be nicer than taking advantage of the new object support in PHP 5, adding the power of XML, washing it down with a natural syntactical (standardized) template engine, and letting the rest come to a head.
For the record, I never thought I would be advocating XML, never thought I would learn XSLT, and most certainly never figured I'd be putting Java at the top of my languages to get good at. First though, I must unlearn the procedural methods taught to me by several years of college.
After the training, I came back to the room briefly before Jason came by. Our destination was a pizza place whose name is actually escaping me at the moment. The pizza was very good, though it was a bit on the pricey side. It was worth it though. We then walked Union, drove around for a while, talked about girls, conventions, and life, and parted ways. Made for a good last night in San Francisco. Tomorrow is day 3 of the training, focusing on aggregated layouts (forced content), some cool tweaks, and feedback. I'll probably type my last entry on the BART ride home.
In response to "Another Day of Training":
Are you there?
I was waiting for this kind of posts ...