Regarding the last item, it seems that Google could have done better regarding tasks management. Mail and Contacts seem very well integrated, but Tasks seem a bit out-of-place. What’s that little pop-up windows all about? I think that’s the only thing that makes me use a “foreign” app in my workflow.

If Tasks were to follow the EverNote or OneNote approach, that would be perfect. And after that… maybe a Google based file syncing service. And them… web-based mind-mapping. And them…

¹ from “The Shawshank Redemption” (@ imdb)

The other day, during “team motivation” training there was an exercise where a colleague tried to motivate me to work with him in a new project. The idea was to have a “coaching” session in order to perceive my expectations, but the way he approached the session made me thing about what I am and what I would like to be doing in a short/medium term.

Having a computer science degree, I’m obviously very interested in Software engineering, but above all I consider myself a Systems Architect/Engineer. Of course I’m able to understand an existing “blue-print” and build the building (give me a specification, an editor and a compiler, and I’ll be a very happy man), but I’m also very interested in participating in the “sketching” of the “blue-print”. Especially, and above all, I like have a global view of what is being built.

So, this colleague suggested a baby-steps approach, where a complete birds-eye view of the system was only achieved after several months. I don’t want to start by implementing very specific I/O routines, or be in charge of a very specific infrastructure component. I don’t believe in having computer guys working the infrastructure (low-level design, I/O, supporting middleware) and leaving all the domain stuff (domain algorithms, specific computations and modelling) to the domain experts. Software should be built by software engineers with the support of domain engineers ([DDD] has a very nice discussion on this issue). I need to see the complete architecture of what I’m building, that’s when I most capable and most efficient.

References:

[DDD] Eric Evans , “Domain Driven Design: Tackling Complexity in the Heart of Business Software”, 2002 (@ google books)

$ mkfs -q /dev/ram1 8192   # 8192 = 8MB disk capacity
$ mkdir -p /mnt/ramcache
$ mount /dev/ram1 /mnt/ramcache
$ df -H | grep ramcache

to mount your own mini ramdisk.

For more details, here and here are some references I found on the subject.

You might ask yourself “Why does this guy needs a ramdisk?… What a geek!”, and wouldn’t mind. It was kind of a epiphany, when I remembered to use a ramdisk to run a “‘Lack of space’ Test Case” on the application I’m building right now.
After running the application, I just had to copy a mp3 album to occupy the remaining disk space… et voilá, welcome the expected “error message”. No, not a SIGSEGV, a real error dialog!…

redirects the stdout from ‘command’ to a file called ‘filename’. The file shall be created if not present, or otherwise overwritten.

> filename
: > filename

are a special case of redirection, creating a zero-lenght file. The ‘:’ is a dummy placeholder that generates no content - usually used when the first form is not allowed by the shell.

command >> filename

also redirects the stdout to file, but the content is appended to file in the file already exists.

note: pre-pending 1, 2 or & will redirect stdout, stderr or both , according to the redirection ‘operator’ (>, >>).

No more, no less… just perfect!
Time to rest.

Andrew Hunt is becoming one of my favorite authors. The first book I read from Andy was “The Pragmatic Programmer“, which is a book full of insights on the life philosophy to become an expert programmer. Now I got my hands in “Pragmatic Thinking & Learning - Refactor Your Wetware” and it is another excellent book. This book is not only for software developers, it is intended for a wider audience and provides lots of tips to improve the way you think - it certainly made me think about the way I put by “mind” into the job.
I strongly recommend it…