Archive for November, 2007


November 29, 2007

Last Friday Nov 16 at Web Develpment Ways meeting by XeDotNet, Davide Senatore introduced us to DotNetNuke.

I’ve never played with DNN, so Davide’s session was really usefull to me;
his speech borns from his day-to-day job so it was everything but an adv spot.

Maybe I’ll have a look deeper in the near future.

Thanks Davide.

Technorati tags: XeDotNet, DotNetNuke

ONE-Day Web Development by XeDotNet

November 25, 2007

OneDayWebDevelopmentXyDotNetXeDotNet is on the road for the first 2008 big event in North-Est of Italy.

Next Jan 17 in Venice (Mestre) a full day meeting about web development is waiting for you:
4 sessions about Silverligh, Ajax, Live ID.

You may have already read on this post in Andrea Boschin‘s blog.

Location: Novotel, in Venice-Mestre (map)
9:30 – La piattaforma Microsoft per il Web
Speaker: Pietro Brambati

10:15 – Certezza dellโ€™identitร  nellโ€™autenticazione web: Windows Live ID come possibile soluzione
Speaker: Davide Vernole

11:40 – Rich Internet Application nel mondo reale: Silverlight e AJAX alleati per migliorare la User Experience
Speaker: Andrea Boschin

13:50 – Applied ASP.NET AJAX 1.0
Speaker: Andrea Dottor

15:00 – Web Games con Silverlight 1.1
Speaker: Davide Senatore

I’m talking face to face to people who live in north-est of Italy:
you’d better not miss this opportunity! ๐Ÿ˜‰

Subscription already open: subscribe now!

Are you still there ?

source (italian lang) :ย XeDotNet

Thanks to Microsoft Italy sponsorship. ๐Ÿ™‚

Technorati tags: XeDotNet, One-Day, Web, Development, Event

Five ways not to fail a project … with a change

November 21, 2007

I found very interesting Joel Spolsky’s How Hard Could It Be?: Five Easy Ways to Fail .

Someone of you may think I write about this argument too often.

Please, follow me one more time:what if your new house’s price rise 10-20% a day after the mortgage contract ?
or house’s deadline will delay by weeks or even months ?

Now think what if that are your money … ok ok ok it’s always customer’s money … only money ๐Ÿ˜‰
So, what abount thinking you are the customer that want the project done by time, by money, and without bugs. ๐Ÿ™‚
Are customers dreaming ?

I don’t know how many software projects get the full score (time, money, bugs), but I know sometime we have to look carefully every single detail.

Five points list by Joel is very interesting, but from my italian experience I’d suggest one change.

Usually, a developer cannot be on a single project for a full 40h/week, every single week, for whole project’s life;
he/she will move among different tasks, be at meetings, etc, of course.
When you evaluate your deadline, please consider developer 100% as 30 hours/week, instead of 40h/week.
Of course every member of the team will work 40h/week, but time spent will be 30h/week, more or less;
that 25% missing represent unplanned new entry, meeting or some very hard bugfixing session; you might consider it some sort of parachute, but it’ll help you to define a more reasonable milestone.

25% !?!?!? @#%&ยฃ!!!!! It’s a huge amount of time, you wrong in some other place for sure.

You’re right, that’s orrible I know, but this is a reasonable value to me so far; maybe I met very very optimistic people ๐Ÿ™‚
or it’s some sort of euristic estimate without too much analisys. ๐Ÿ˜€

disclaimer: this is based on my personal experience. ๐Ÿ˜‰

Technorati tags: Planning, Estimating

Cell BE and PS3 bookmarks

November 14, 2007

Whatever …
recently I spent my time learning parallel computing, clusters, grids and so on.
I found very interesting latest news about PS3‘s clusters, so I decide to spend more time and have deeper look.

PS3’s heart is cpu CELL Broadband Engine, or simply Cell.

I won’t talk about Cell‘s logic, the PPE unit, 8 SPUs (SIMD processors), 204.8 Gflop/s theorical peak performance or EIB bus (that moves 96 bytes every single cycle!).

Well, this little toy is the engine inside PS3 and someone had the right idea to “play” with it to build up a sort of cluster.

Unfortunately main goal of PS3 is not number cranching, massive calculation otherwise we’ll see serverfarms full of PS3. ๐Ÿ™‚
That would be lots of money, STI’s dream (Sony-Toshiba-Ibm) would come true. ๐Ÿ™‚
But (there is always a but!) PS3 is powerfull enough to solve small problems, without spending lots of money.

PS3 Gravity Grid
-IBM ray tracing and youtube video (plus a pdf here)

Impressive, isn’t it?

If you want look inside the box, here are some docs about Cell cpu:
Cell Programmin Handbook
IBM Cell resource center
IBM Mathematical Acceleration Subsystem
-SCOP3, A Rough Guide to Scientific Computing On the PlayStation 3
Build an 8 PS3 supercomputer

As you’ve seen above, this is not rocket science but it may help you to walk your way easily enough.

Technorati tags: Cell BE, PS3, Grid Computing, Parallel Computing, Distributed Computing

Getting the goal = Planning & Estimating

November 7, 2007

Then I come back to an important issue: getting the project to the goal.

Everybody met a bad project at least once in his life;
everyone of us reacted in different ways: developers, testers, architects, management … and so on, well you know!

When you face up such projects, you learn something that let you understand it next time earlier.
Since “nobody already knows future” (funny italian: nessuno nasce imparato ๐Ÿ™‚ ) other people experience help us improving our sixt sense. ๐Ÿ˜‰

Joel’s Evidence Based Scheduling article tell us about his way to manage a project:
-keep your tasklist detailed
-keep memory of time scheduled
-keep memory of time used
-keep memory of extra unplanned time (bugs fixing, meeting, new featues, etc)
… mix it up all together, play some math operations and you’ll have an almost exact scheduled task list.

Well, it’s not exactly this way ๐Ÿ˜‰
It’s much much more simpler.
You’d better read it straight from the source .

Of course, it helps you even if you haven’t already met that type of probject.

forget about it! ๐Ÿ˜‰

Technorati tags: Planning