28 April 2009

Code as Exploration of Unknown Territory

I have come to view coding as the act of exploring and charting unknown territory -- the wilderness of our cognitive space.

If we let it, as we go along, our code illuminates the crevices and crags of our understanding, and shapes as it goes, our ideas of where to go next.

Version-control is the key mapping tool.

I've just reached a point in the development of the mikroblogging tool where I (finally!) believe I understand what's needed; what might work as something new-ish and interesting-ish in the conversations we have over this Internet thingie...

It's not a conventional website like Twitter or Blogger. It's not a standalone desktop/PC-installed system doing some sort of smart p2p stuff. It's not a conversation follower nor a search tool nor a Bayesian-interest-detector. More something hybrid from all of these.

Let's see where the code will take me next...

21 April 2009

Sun and Oracle

TL;DR: A fucking disaster for everybody except Oracle and Sun's execs and (maybe) shareholders. i.e. The Kakistocracy wins and the rest of us get shafted. (As usual.)

I think Oracle are getting an absolute bargain. $7.4 billion is chump-change for the IP they're acquiring. The question is, Which pieces are they going to keep, and which are toast? Some of this is blatantly obvious, other bits are pure crystal-ball-gazing on my part.

Java: the no-brainer. Oracle are heavily invested in Java for the Enterprise stack. Bad news for open-source Java? Maybe! Or maybe not... it all depends on whether Oracle were backing Apache simply as a tactic against Sun in the JCP. I've heard a number comments along the lines of "it's GPL -- the boat has sailed". They're forgetting that the owner of the IP can do as they please, including closing the source completely in the next release. (Not that I think it very likely; just a possibility.) Yes, an open-source community might be able to follow, but I'm betting that there would be compatibility problems.

MySQL: Toast. (Personally I never bought the logic behind Sun acquiring MySQL, and then they went and mishandled the whole thing badly.)

Glassfish: More toast. What will this be... Oracles fourth appserver? I lose count.

Netbeans: A cold shaft of ice pierces my gut. I love Netbeans. I just don't much like its competitor (just a personal preference; don't read too much into it!) But I fear that this might be the end of the road for NB... OTOH it can -- unlike so many of Sun's other open-source projects -- probably survive, nay flourish, as a standalone open-source IDE. After all, that where it came from in the first place.

OpenOffice.org: Makes perfect sense for Oracle. I bet on them keeping this one going. In fact, this might be the Secret Weapon Acquisition... the knife with which Oracle goes seriously for Microsoft's jugular in the Enterprise space, together with Solaris and the Sun hardware.

Solaris: I'm betting it stays. Oracle's strategy has been (to the limited extent I bother to keep track) to lock up the mission-critical, "hard-to-do" stuff in the Enterprise space. And there's still a whole lot of stuff that Solaris does way better than Linux.

VirtualBox: also plays well into the Enterprise/datacentre "integrated offering" strategy.

JavaFX: not one I'm capable of guessing about... any offers?

And the Sun hardware: makes pretty good sense for Oracle, in my limited understanding.


On a more personal note, much closer to home: How likely is it that Oracle will retain Sun's training programs for Java? Methinks it unlikely, since they already have their own training programs. What does that mean for the first Sun-authorised Java trainer in Africa?

Not good news at all, I'll bet!

All-in-all I think the acquisition is terrible news for Sun people, and probably not good news for their customers, either. I am finding it hard to see it in a good light for Java, either, and, having (literally) bet the farm on Java for the past 13 years, find the prospects quite discouraging. And for open-source in general it's a disaster. Despite the Slashdot whiners, Sun has sunk an incredible amount of money and effort into open-source projects, and I simply don't believe that Oracle has the same largeness of vision.


Oh well. Shit happens. I suppose its still a step better than Sun going under completely... Best I get a move-on with further developing my own training material and courseware.

13 April 2009

Twitter

Twitter looks vaguely interesting. Not too much. Not enough for me to bother with it. I think it's very Flavour Of The Day.

But it does suit one thing I've had in mind for a while... the idea of a "stream of consciousness" blog sort of thing. Essentially a blog where I can just post a line or two or three, without all the formality and palaver of Subject lines, Categories, Tags, etc. In a nutshell, Twitter,but without the 140 character limit, and hosted on my own server as part of my own infrastructure.

Maybe I'll just write it....

<i>I need another development project like I need more holes in my head. Only two major development projects on the go at the moment, and a couple of minor ones.</i>

26 March 2009

SL-275 and Java Programmer Certification

Midweek; past the halfway mark (in time, anyway) of teaching Sun's SL-275 "Java Programming Language" course. I wonder how many times I've taught this course over the past 12 years... I still love it! Even though its just the basics, there's something just plain fun about introducing Java to new minds.

This one is interesting; 5 of the students have flown from Stockholm, Sweden all the way to Cape Town, especially to attend this course! Wow! (Of course they might pick up some sunshine in the middle of the Northern Winter. I guess that may help. ;-)

The trouble is that Java has grown so large over the years that there's a hard choice: cover (mostly) everything in a shallow way? or leave a lot out to get more depth?

Add to that, Sun punt the course as "the" prerequisite for writing the Java Programmer Certification Exam. Frankly, I shudder at the thought! Programmers would need a lot more than just this course to be in a position to tackle the exam. And programmers who are in a good position to tackle the exam don't really need this course!

As it currently stands, the course is inadequate for students attending it in order to prepare for the Certification Exam. There's very little I can teach them and still remain reasonably close to the course materials. The course is pretty good as an intro to Java for experienced developers, but pressure from more experienced developers in the class -- especially when they are in a significant majority -- could, if I'm not careful and strong in controlling it, leave the less-well-versed-in-Java students stranded. (I assure you that I don't let this happen.)

The solution is to split the course in two. One would be a truly introductory course aimed at otherwise-experienced developers. And the second would avoid all basic material -- syntax of if statements, declaring classes and interfaces -- in favour of homing-in on the deeper, less well known details of Java execution that the Certification Exam aims to test.

I sure wish Sun would do it soon! It would make life so much easier for us trainers, and would deliver a much better focussed value to customers.

18 January 2009

Open Course Development

It strikes me that, whilst I am deep into the design of a number of programming-related courses, I am making a terrible mistake. The mistake of not talking about what I'm up to. The mistake of assuming the entire burden of course development.

Instead I ought to be employing Open Source development principles. At least some of them. Seems to me that (at least) the Thousand Eyeballs principle applies quite well, at least to the course structure, content and sequencing.

Right now I'm in the thick of codesigning three courses that I (with good reason!) believe are served very poorly by the corporate training world, and I think I can design courses that deliver much, much better value for money. The three I'm busy with right now are "Elements of Object Oriented Programming", "OO Analysis and Design" and "Patterns of Software Design".

"Elements" is intended as an introductory foundation course for programmers coming from a non-OO background who want/need to learn the OO concepts quite quickly. Nobody in their right minds would believe that a 3-day "Elements"-style course is going to turn any Natural programmer into an OO expert... we all know that the process of learning to Think Objects takes 9 months. But I do believe that it is possible to teach the foundation concepts really well.

"OO A&D" seeks to demystify the analysis and design process -- to the degree that it is possible to demystify an essentially creative process.

And, finally, the "Design Patterns" course aims to teach something useful about... well... design patterns. I've taught variations on this one numerous times over the years, but never really been satisfied with the value I was able to deliver, so I've completely rethought the approach from the ground up, and -- I think -- come up with something Just a Bit Different. A whole lot of the Design Patterns courses I've seen offered seem to add up to nothing more than a whole lot of droning through the GoF patterns catalogue. I think there's a whole bunch more to the topic than that, and I plan to use the course to explore that.

I've also outlined a whole bunch of other courses that I want to develop over the next year or so, but developing even these three is a daunting enough task for now.

I'm openly stealing some of the concepts -- the approach to teaching -- exemplified by the Head First books, and I'm really interested to see that O'Reilly are themselves developing a bunch of courses -- I gather they'll be web-based courses -- under the Head First brand. Good for them! It's about time we saw a better approach to teaching technical stuff.

So I'm aiming to make the courses colourful, interactive and fun. I'm trying to build in lots of of pictures, music, video, games, practical exercises, movement. (Trying to figure out how to include flavours and smells... ;-)

I've also been breaking away -- especially for the "Elements" course -- from linearity in the course sequencing. Essentially I'm developing using a Spiral Model. First introduce a concept, then go on to related concepts, etc., in time spiralling round to repeat the discussion of the topic in greater depth, and so on. I've long known that it is too easy to give too much detail all at once, so for my own course material, I'm explicitly shying away from that. I think it equates -- somewhat -- to the concept of Progressive Disclosure in user-interface design. As luck would have it, after several weeks working on this stuff, I tripped across this article just today, and went "Aha! Somebody else who Gets It!"

Right now I am stuck. Struggling to come up with good great practical ways of exploring the topic of "Encapsulation" in a way that doesn't stray too far from the way it's meant in OO programming, whilst remaining vaguely interesting.

As soon as I get the relevant bits of machinery up and running, I'll post the course outlines. (Will take a while: I am unexpectedly and suddenly off to London for a week for a spot of consulting work.)

30 December 2008

Teaching Servlets and JSP

It's the end of a week teaching Sun's SL-314 course -- "Web Component Development with Servlets and JSP" in Cape Town. It's been challenging and demanding, but, above all, fun! I had a great bunch of participants on the course, a couple of whom were old friends from an "Intro to Java" course (Sun's SL-275 course) some months ago.

The course materials are pretty good, although with a couple of misfeatures -- in my ever-so-humble opinion, as always -- and one glaring omission! (Never mind the details; I won't bite the hand that feeds me. This time. Rest assured that I've passed my thoughts on to the good people at Sun Education.) However, the whole sequencing and structure of the course has set me thinking, "How would I go about structuring such a course? What examples and challenges would I use for lab work?"

I take this teaching stuff pretty seriously1. So I'm always looking for ways to add value above and beyond what is offered by the course materials. Of course, the most valuable thing a trainer can bring into these sorts of skills-oriented technical courses is one's own experience actually using the technology on real-world projects. But beyond that, a lot of value can be added (or subtracted) by the sequence, timing and manner in which the material is presented. In a 5-day course time is terribly limited, so practical examples and labwork are necessarily very constrained, so it is very hard to work-up really good examples.

Setting examples and lab-work for courses is actually a very hard problem. You don't want to restrict the work to toy problems -- students pick up on that very quickly, and, even more quickly, lose interest in the course material, and, more importantly and with longer-term consequences for the remainder of the course, they lose respect for the course (and sometimes for the instructor, too.) On the other hand, time is very limited -- perhaps a couple of hours for any given issue, so extensive examples take too long to implement.

Then, too, one is in the position of trying to illustrate and/or reinforce key points in the course material, and developing actual code always involves a measure of extraneous detail -- housekeeping code. It adds nothing to the teaching purpose of the exercise, but it certainly chews into the time available!

And -- not to be neglected -- I think it is important that examples be at least a little bit entertaining. After all, we can do order-entry at work anytime... But there's a borderline across which "entertaining" becomes "irrelevant". (Much the same can be said of micro-benchmarks.)

Relevance? You wanted relevance from a blog post? As opposed to rambling?

OK.

I'm starting to put together my own material for a number of courses, centring around the gaps in the corporate professional training space (OO and Java technical matters, of course.) that I've been seeing over the past decade.

(Of course, now is a brilliant time to be putting my own course-material together. In times of economic downturn, the very first thing to get the axe is training! How short-sighted is that?)


[1] As with most things, the more you put into it, the more you get out of it!

15 November 2008

Brain Delayed - In Transit

On my way down to Cape Town -- yet again -- for a Java Intro course -- yet again...

On arrival at the checking-desk, get told that I'm not booked to go to Cape Town, but am eventually found to be on my way to George from Cape Town. Returning to Cape Town on Friday. Wow! What happened here? OK, so off to the booking office where I have to fork out an extra R700 to cover the ticket change. Fortunately the SAA staff at George and I know each other by first name. (As a result I seldom have to produce ID to check in -- most of the staff can ID me by sight. Beat that for personal service! ;-)

Sudden thought: Let's check the car rental reservation. Sure enough! I'm booked to pick up a car in George. Get that one changed. Things are so screwed up at this point that I called the B&B where I'm allegedly booked to stay in CT. Yes, indeed, they DO have a reservation for me.

WTF was somebody at the travel agency thinking? This guy is going to fly from Cape Town to George, pick up a car there, and then stay 5 nights in Cape Town before reversing the journey?

Now I only have to waste an extra hour-and-a-half in the airport. I just love hanging about in airports, don't you?
Related Posts Plugin for WordPress, Blogger...