The Thinkpad has finally arrived and it’s very cool. So far I’m not having any of the issues I had with the Dell:
- Wireless card works
- Keyboard is very nice
- It’s only slightly thicker, wider and heavier than my other Thinkpad (a T40)
- Screen is fantastic even at the maximum resolution: 1600 x 1200
- It didn’t come filled with crap, but only the software I needed. This will be my chance to try OpenOffice ;)
- Oh, and it’s black!
A while back I started putting together a tutorial for installing what I called WAMP (I thought I was being so smart). It was LAMP (Linux-Apache-MySQL-PHP) with the L part replaced with a W for Windows.
As it turns out, not only it’s a widely used term, there is also a much better and simpler way. Just download and install the all-in-one, well, WAMP from here.
The only reason I found this out is that I was trying to install a “Wikipedia”-like server on my machine. Check it out.
I missed last week’s “Silicon Valley Patterns Group” meeting and the week before that it had been cancelled because of other conflicts. I was getting geek withdrawals and I was glad to finally meet again yesterday.
We’re still studying SICP but we’re getting close to the end. Last night we beat streams to death again, which is great, since I had missed the meeting :) Apparently the week before the group had struggled with infinite loops in part because they had missed a certain footnote in the book, which warned against special form expressions.
Some of us had Java implementations of the sieve, with more or less complex solutions.
It’s likely that next time we will be talking about the metacircular evaluator in Scheme, which is the climax of the whole SICP thing. Then we need to seriously start thinking about what to learn next. It seems that it’s going to be another language, and so far the candidates are: Haskell, Common Lisp, Prolog and ML.
Can’t wait.
Ed Burnette took a bunch of pictures at the Jazz demo the other day, during EclipseCon. Here and here.
Jazz is hands down, the best Eclipse app I’ve seen. It has the potential for revolutionizing the IDE space, like IntelliJ Idea did 6 years ago (some would even say “like *Eclipse* did”, but 6 years ago Eclipse wasn’t anywhere compared to IntelliJ, so I wouldn’t push that far ;)
Speaking of pictures, I’m glad I didn’t bother carrying my camera around. Ed took pictures all the pictures of EclipseCon 06 I can possibly want. I’m glad he’s sharing them too.
If you’re seeing a broken welcome page in Eclipse (any version), you have IE7 Beta 1 or 2 installed. I finally discovered this, was driving me nuts.
The bug was filed here.
More games than me (and by more, I mean way more)
No picture because:
[Note: Be a pal and please link to this article, not directly to the images. Thanks!]
After posting about the terseness of Python vs Java (and C and Cobol) I was reminded that new versions of Cobol aren’t as verbose. In fact you can even write “Hello World” in one line of code. This makes Java, and even C, more verbose than Cobol.
Sure, “HelloWorld” isn’t really a good indicator of terseness anyway, and it’s really meant as a curiosity and for poking fun. Though I won’t use it to make fun of Cobol again, since it’s no longer funny :)
The “Shortest Wiki Contest” is not a bad way to get a feel for terseness. Ward Cunningham had mentioned the contest at EclipseCon, though he suggested the winner was “5 lines of Python”. I couldn’t find that solution, the winner seems to be a 4 line Perl script. Nice. Maybe Ward meant “the shortest solution that can still be read by a human”.
But the problem with the contest is that most solutions are implemented with the usual suspects: Perl, Ruby and Python and you can’t really tell how others fare. Though the Java/JSP version can be found here.
As for me, I don’t cease to be impressed by how terse Python is when compared to Java and C.
Dear Customer,
Thank you for choosing Lenovo!
All or part of your order has shipped. [...]
Two weeks ahead of schedule. Go them.
Greg Stein came by the blog last night to say hi, hi Greg! :p . Greg is an avid gamer and when he’s not playing he’s probably busy being the Apache Foundation Chairman, or doing stuff at Google, or whatever.
And today Wired is running an interesting story. A World of Warcraft gamer who got a manager level position at Yahoo, in part because (and not in spite of) he plays WoW.
Next time your SO or your parents, or even your kids, tell you you’re wasting your life to some pixels, you’ll know what to say ;)
http://blogs.zdnet.com/Burnette/?p=51
In addition to winning the award for best Open Source Eclipse developer tool, the last time I talked to them they had no less than *5* job offers while at the conference.
Well done.
Today was a mixed bag. I missed Tod Nielsen’d keynote (Borland’s CEO) and started with the TPTP session. Not what I had planned, but it worked out okay. TPTP is something I’ve been looking into and I find it quite interesting. I was avoiding going to those sessions, because I knew I would have to add yet one more thing to my list of “things I really need to spend some time looking into”. Sure enough, it happened.
Then I did go a talk on how to use and extend WTP, or more specifically, a subset of the Web Tools Project, the J2EE Standard Tools (JST). It was not very interesting. The actual material seemed interesting, and it probably is, but the person speaking could put to sleep any ADHD person on a strict caffeine diet.
I attented some short talks (9 minutes each) on all sorts of different Eclipse related topics. Not so bad, very focused.
Then I closed with the 3 folks who wrote RadRails. Impressive kids. I wonder if the WTP guy could put them to sleep.
Overall for the 4 days, I’d give EclipseCon ‘06 a 4/5 stars. If I had my own company, I would hope that my developers would want to go to these kinds of conferences. They’re great for regenerating brain cells.
With EclipseCon and everything, I haven’t spent much time with The Python Challenge and I’m still at level 5. But here are my solutions thus far.
1) Ward Cunningham gave an interesting talk on collaboration and trust. On a different thread, he said that the first wiki he wrote was done in 400 lines of code which he then improved to 200 lines. There is/was a wiki contest to write the smallest possible version and the winner was a 5 line program in Python. I thought that was interesting.
2) A panel kind of session, with folks from several companies, answering questions about RCP.
- At the end of this session, I made fun of Jeff after he asked someone from SAS if he knew some person X. Like SAS is some 12 people company.
3) Jeff gave an interesting (though quick) demo on how to build a headless Eclipse plugin
- After the talk, Jeff made fun of me after he asked yet another person from SAS if he knew person X, and this person knew him.
I’ve been following all the sessions I can possibly attend, take notes, read more material when it’s relevant, etc. But today I’ve decided to do an experiment. I was going to sacrifice a talk by doing things on the computer, while I was attending the session.
It had to be a session I would normally be interested in (or it would not be a fair experiment) and it also had to be a subject that was not hugely important and there’s nothing better at the same time (or it would not be fair to me). I figured that the talk on writing Eclipse processors for annotations, was a good candidate. It’s something I’m interested in, but if I miss it, I won’t be too sad. This is a 45 minute talk too.
There I was, trying to listen and take notes at the same time, doing email, responding to IM, and other stuff people usually do in meetings.
I can honestly say that for the most part, I have no idea what these people talked about. The subject didn’t seem too complicated, certainly less than EMF for example. Actually 5 minutes in the talk, and I was already missing a lot of context.
Lo and behold, I’ve made this enormous discovery: If you don’t pay attention, in addition to annoying other folks with your continuous click-click (I saw their looks) you won’t be able to follow what’s happening.
Don’t do it! Unless of course, you’re doing your own antisocial experiment ;)
I almost didn’t go to Greg Stein’s keynote (figured I would learn more if I stayed in bed an extra hour) but I’m really glad I attended. Greg (the current Apache Foundation chairman) gave a talk mostly on Apache, though he tried to compare it to Eclipse whenever it made sense. He basically talked about licenses and open source, and compared the freedom of Apache Licenses (more freedom for developers) to the Eclipse License (more freedom to end-users).
He’s convinced that in another 5 to 10 years, there will be no money to be made in packaged software. Though there will be plenty of money in services. The exceptions would be products that require a lot of infrastructure, such as Google (he’s at Google now) or games and their masive amounts of (expensive) content.
Other than being a good speaker and appearing smart (he was on stage, looked tall, so I’m sure he’s smart), he also looked like a genuinely nice guy.
Not a funny talk like Joel’s, but good stuff. Definitely worth losing an hour sleep.
After that I went see a demo of an IBM Business Intelligence tool. Mostly a demonstration of “all these Eclipse technologies we leverage”, such as EMF, GEF, DTP, GMF. It was nice.
I’m sitting here in the lobby, waiting for the next talk, on annotations, and something just occurred to me. I look at all these people walking by, and some faces are very familiar to me. But I don’t know if I know them and they know me, or if I just saw them talk, or if it’s just the 5th time I see them today or what. Since I don’t know if I know them or not, I’m just going to go with the latter, and try to get inside the monitor when that happens.
Greg Stein just walked by, I’m pretty sure I should not wave and smile.
Anyway, time to go see people annotate stuff.
Today was a great Eclipse day.
Joel Spolsky opened the conference and what a show. His talk had nothing to do with Eclipse, or did it? Joel talked about design and what makes a desirable product vs. the rest. He compared iPods to other m3 players, Aeron chairs to clones and even Muzak to Lynyrd Skynyrd’s Sweet Old Alabama.
His point was that isn’t always obvious to guess why a product makes it and others don’t, so we came up with “The Formula ™”. Joel says that after a make a functional product you can make people happy, work on the emtional side of things and make it pretty, you’re pretty much gold. Lots of features and such, is not where the trick is.
He had the best PowerPoint presentation I’ve ever seen, a truly work of art.
Marvellous stuff, I’m glad I was right there in the front row.
Then I went to a “Gentle presentation to GEF” (Graphical Eclipse Framework) and the conclusion was that there is no gentle way to introduce GEF. In a nutshell, GEF is a framework for building rich editors based on a model. Really neat stuff, but like I said, steep learning curve.
From there I went to see what Brian Payton was up to in the Data Tools Project. Data Tools, or DTP, is the series of Eclipse plug-ins for helping with database applications. It comes with connectivity plug-ins, parsers, SQL builders, models, etc. Good stuff, even though it’s only in its 0.7 version. 1.0 in June, with the rest of Callisto.
Erich Gamma (Mr Design Patterns) and John Wiegland were next. They talked about some Eclipse metrics, how bugs flow from project to project, and from component to component within the same project. Quite interesting stuff, and the app to visualize it had a very interesting UI.
They also talked about the development process Eclipse uses, which is a mix of Extreme Programming, Scrum and other agile practices. It works for big remote teams, as I would expect.
The next presentation I went to was on GMF, which is a layer on top of GEF and EMF. If you know you’re going to use EMF and GEF, go with GMF and save huge amounts of time. I’m definitely planning on learning EMF and GEF after I’ve played with GMF, not before. A different opinion here.
And finally for the biggest wow of the day. The folks at IBM, and more specifically the team led by Erich Gamma (again), announced the project they’ve been working on for the last year. A client and server development team collaboration tool on top of Eclipse. They’ve demo’ed a real life situation, with one team lead and two developers working simultaneously. Not only the UI is really sweet, the best I’ve seen on any Eclipse-based application, but the integration between different development tools to make a really seamless workflow, is going to be a killer. Can’t wait to get a hold of it.
5 stars for today, no contest.
Jeff had a nice article last week, about performance reviews. It linked to another article that talked about how annual reviews should focus on judging potential, rather than compensating for work that was already done. Just like job inteviews. I responded to Jeff to basically say that I agreed with him, and that I am frustrated to see that none of that good stuff is happening. Judging from what has been going on in the last 30-40 years, it probably won’t change anytime soon.
I was going to clarify my response but rather than doing it in the comments section, since it was becoming its own essay, I chose to bring it to the top of the page (great, now I can really have a conversation with myself, and respond to my own posts)
–
So why aren’t performance reviews based on potential? Why can’t we do it, since we manage to it when hiring someone? And with less information too!
It’s done while hiring only because there’s no easier alternative. If there were, it would be the technique of choice for hiring people. Moreover, we can stand this lack of alternatives, because we can shift most of the blame and share what’s left of it with other folks.
1) Shifting blame: “Who would have thunk, he fooled us during the interviews! Our process is bad, let’s revisit it”
2) Diluting blame: Have lots of folks interview the candidate. That way if the group makes a mistake, if will be a collective mistake. Nothing that can’t be fixed with a group hug.
When it’s time for yearly reviews, most managers feel they don’t have that luxury (shifting and diluting blame). They are now the sole interviewer. Any mistake, they’ll be the only ones to blame. No more poor excuses and more importantly, no group with whom to share the blame with.
It’s too risky to make a mistake, and on top of that, it’s also hard to reach a good conclusion. So what’s the poor manager to do?
The obvious approach is to make it easier to reach the decision. We trade judging potential (too hard and risky) for compensating past performance (less risky). Compensating past performance is indeed less risky, but it’s still hard, so what can we do that’s both risk-free and easy to do?
I believe the winner approach would be: Judging Effort.
Effort is something you can’t deny. Everyone can see it has happened, no need to justify it too hard. It can be easily measured: more lines of code, more hours in the office, more meetings called and attended, more projects started, etc.
Whereas the original article advocates “hard and risky”, I was pointing out that (unfortunately) it won’t happen, because we humans are what we are. We will pick “easy and safe” over “hard and risky” every chance we get. And that’s irrespective of the potential of the answer (there’s a meta-evalutation thing going on here :-)
