One of the complaints I keep hearing is that with the command line, people have to type in name of the executable to open a document. Whereas in the GUI world, you just double-click on a document, or on the application icon. And I realized it was not so much about typing the application name, it was more about having to remember application executable name along with the file name.

So here is a solution for those of you on Linux and possibly other Unix systems. What we need is an application which can identify the application for a specific file type and open it for us. And we already have something for this – xdg-open. It is part of the XdgUtils toolkit which aims at better desktop integration across various desktop environments. And most of the popular DEs qualify to be XDG compliant.

xdg-open opens a file, including URLs, using the preferred applications. In short, you can type xdg-open resume.odt and it opens up OpenOffice, or your preferred application, for you to edit the document. If you find xdg-open to be cumbersome to type, create a symbolic link with your favourite command. For example, if I want to open files using the command o, I do

ln -s /usr/bin/xdg-open /usr/local/bin/o

Now I type o resume.odt or o blog.py or o http://ifacethoughts.net to open any of the corresponding applications and work on those files.

We can perhaps take this one step further by using command_not_found extension for bash to invoke the application by just typing in the file name, without having to worry about the application name at all.

I personally think that the best way for patents to be implemented for software, is the open source way. If a patent is about exclusive rights in exchange for public disclosure of an invention, there is no way to claim a patent other than open sourcing the software invention.

The problem with patents in software, and maybe not in other fields, is that it harms everyone involved. When reverse engineering works with most of the software, public disclosure is not necessary nor useful. Secondly, it is extremely easy to copy, whether it is the design or the implementation (which I consider a benefit) in the software world. And most importantly there are thousands of ways of doing the same thing here. In such a system patents turn out to be weak supports. And companies that resort to such a weak system become unpopular with the masses.

A better way is to allow people to copy it, but with due credit and without violating your terms. And this is open source. This has proven well, it has helped speed up adoption and evolution of ideas and also improve implementations.

I can feel it when John Carmack says:

The idea that I can be presented with a problem, set out to logically solve it with the tools at hand, and wind up with a program that could not be legally used because someone else followed the same logical steps some years ago and filed for a patent on it is horrifying.

Maybe patents have been the cause of innovations elsewhere, but in software they have mostly been the cause of problems.

Alain Ranaud presents a fresh perspective in support of software patents, but in my opinion, it still highlights the problems with the patent system than its benefits.

If we apply this logic, every recommendation will be against some competitor. If the Government recommends Microsoft platform, the Linux or Mac guys can raise a hell. Just like a proprietary company might do it now if a country proposes the Linux platform. Or if a country prefers Web applications then desktop companies might revolt. Right now this is well marinated with a paste of anti-capitalism or anti-IP or a similar flavour. But we know that open source is an effective way to be most strict about IP. I am no economist, but I that open source is anti-monopolism, and not anti-capitalism, because that works the best for the user and his/her freedom.

To think of it, this is not surprising. FUD, and not merit, has been the most-used weapon against open source. And this is just another instance of it.

The motive behind moving to Arch Linux from Ubuntu was to explore under the hood operations. Arch Linux is more of a meta distribution using which you can build your own. There are options for each and every single component of your OS, and you get a chance to dive deep into it. While all this is possible in Ubuntu as well, Arch Linux makes this the way of using Linux. And it has immensely helped me understand and solve some problems even for other distributions.

My aim of moving back to Ubuntu is two-fold. One – I want to be in the same zone as those whom I recommend using Linux, which is usually Ubuntu. Arch Linux is not for those who want to just use, and not learn, Linux. Also, Ubuntu today is more than an OS, it is an infrastructure solution with the Server Edition, ease of working with utilities like apt-cacher and the cloud services. Two – I think that a combination of Ubuntu and the Arch Linux approach might be the right recipe for many.

If you’re creating a new product, what are the three (or fewer) key features that will make it so great that you can cut or half-ass everything else? Are you focusing at least 80% of your effort on getting those three things right?

Very nicely and concisely put. A must read, even if you are only a user and not a designer.

Both Gmail and iPod followed this advice and managed to revolutionize how things worked.

However, I see some differences between Gmail and iPod. Gmail has continued to evolve and more importantly free the user. The Web interface works with most of the modern browsers on all operating systems. Standard POP, IMAP access works great with many desktop or mobile email clients. I can’t say the same about the iPod. It is inherently biased towards Apple – an implicit vendor lock-in.

In my opinion a product can build a long-term strategy by following Paul’s advice in such a way that it retains the user’s freedom of choice.

I doubt. We were able to get 55MB of RAM usage with plain Openbox and some XFCE accessories on a stock Ubuntu installation (so that the machine could still use the local repository). Even with Abiword or Gnumeric it did not exceed 90MB, and that worked quite swiftly on a machine with 256 MB RAM and old enough to be low-powered than today’s netbooks. All this with font anti-aliasing and pretty window decorations.

So I think that this decision is either because of the trend of choosing browsered applications over desktop applications, or because netbook users do not see office suites as necessary.

Many argue that Web design is no different than other art streams, like painting. I think art plays a big role in Web design, but they are not the same. An exercise of design indicates intention of use, which is addressed better through engineering.

• xmonad: A wonderful tiling window manager that has found a place in all my monitors, with different resolutions
• dmenu: IMHO the best launcher out there
• centerim: Console or not, this is the best and least intrusive way of chatting
• GNU Screen: An immensely useful terminal multiplexer
• git: A version control system that lets you code first and worry about version control later
• Chromium: It just sped past all the existing browsers.

Some others are sitting in my experimentation lab:

• uzbl: A Webkit based browser, which has changed my perception of a Web browser
• zsh: A powerful shell, though I am still finding out what I stand to lose if I move away from bash.

Which ones have you picked up?

It is commendable that Microsoft is complying with the Europen Commission policy and is doing its job. However, as a Web community, we face a bigger challenge in educating the users and corporates who do not care about getting rid of IE-specific applications or working with a better browser. Unless this happens, a free browser choice won’t have much effect.