iface thoughts

XHTML 2 To Be Aborted

Abhijit Nadgouda — Sun, 05 Jul 2009 03:29:23 +0000

W3C has decided to abort the XHTML 2 effort. It is said that this will allow more time and effort towards HTML 5. There were some indications towards lack of interest in the XHTML evolution, so this is not exactly a surprise. XHTML’s strictness made it unpopular with the pragmatists, but I think it lost out more because of popularity of HTML 5. I am sure it will not be missed by the majority, but some of us did wonder if it could have led to more innovative uses. An important thing to mention is that the XML serialization of HTML will be addressed by HTML 5, what we will lose out on is the XHTML family of documents.

Probems Or Symptoms?

Abhijit Nadgouda — Sat, 27 Jun 2009 07:40:09 +0000

It has often happened with me that what I considered was a problem was only a symptom of the real one. For the user, the symptom is usually the problem itself. That is why it is all the more important for a problem solver to sift through the symptoms and arrive at the underlying problem. One of the differences between the symptom and a problem is that the symptom will usually be more about the environmental factors and external systems. This includes those problems that occur intermittently are not easily reproducible. Instead of giving up on them, acknowledging a deeper problem and investigating a wider area might lead us to the real problem.

Psychology in Design

Abhijit Nadgouda — Tue, 23 Jun 2009 02:56:31 +0000

Jeremy Keith has great inputs on designing with psychology in mind. Though he talks primarily about Web designers, it holds true for most of the software application designers as well, or anyone who gets a chance to create an environment. A lot of times designers limit their exercises to observing and studying their users and design accordingly. What they miss is how do they want to change that behaviour through their design to get the desired effect, whether it is better productivity or higher revenue.

Optimizing Code For Reading

Abhijit Nadgouda — Tue, 16 Jun 2009 03:38:01 +0000

Juergen Brendel’s advice about read-optimizing your code is useful. Developers’ love-hate relationship with documentation spans a wide spectrum. You will meet developers who look at commenting as an effect of poor programming and also those who think that all code should be documented. Factors that I use to decide the level of documentation are the domain and the audience for my code. In-line comments can be effectively used for things that cannot be conveyed by code at times - the rationale behind design decisions and the problem the code is trying to solve.

Web2.0 Is Now English

Abhijit Nadgouda — Fri, 12 Jun 2009 02:34:46 +0000

The one millionth word in English dictionary is Web2.0, defined as “the next generation of web products and services, coming soon to a browser near you”. The unfortunate part is soon that next generation is soon going to move on to something else, and continue further. The phrase Web 2.0 itself has been confusing. Its popularity is born more out of marketing than anything else. It was already bad enough that it was exposed to the people, and I think it is a disaster that it is part of the English dictionary now.

Programming And Technology Bylaws

Abhijit Nadgouda — Fri, 05 Jun 2009 18:18:24 +0000

Bill presents a nice list of programming and technology bylaws. Many feel that these are not effective in the practical world. But I see their proof almost everywhere. For example, almost every conversation with a client turns out to be an ideal example of Amara’s law. A good read, and at times a good guide of the non-engineering aspect of programming and technology.

uzbl - The Stripped Down Browser

Abhijit Nadgouda — Mon, 01 Jun 2009 10:41:53 +0000

I am trying out the uzbl browser for last couple of days. It follows the unix philosophy to strip down a browser of everything else other than visiting the web sites. Everything else, even the management of cookies, bookmarks, history and downloads is outsourced to external scripts. It departs radically from the other browsers by using vim-like modes to make it keyboard heavy. So much so that you will not find any menus, icons or buttons. I have been using vimperator for a while, but with uzbl the absence of features has helped it gain a lot of speed and agility.

I also think that this outsourcing model will make uzbl more programmable and reusable. I quickly hacked a one-liner script to post to twitter.

#!/bin/bash

curl –basic –user “:” –data-ascii “status=$8 $6″ “http://twitter.com/statuses/update.json”

This script will post your message to twitter with link of the current web page you are surfing. I have bound this script to the command \tw in the configuration file. Now I can type \tw "my twitter message" in the command mode to post to twitter. This is because uzbl exports certain variables to the scripts so that the browser can be extended.

Another example,

#!/bin/bash

curl “https://:@api.del.icio.us/v1/posts/add?&url=$6&description=$7&tags=$8&replace=no”

Now I use \dlcs "my tags" to bookmark the current page on delicious.

uzbl is still in alpha. It might be too early to say whether it will be popular or not, but I think uzbl will make a big difference by its outsourcing model. You can quickly build your own scripts, use context given by uzbl and do whatever you want to do with it. This is like building Firefox extensions, but without having to learn its framework or technology.

uzbl is not even alpha yet, but it is already a very extensible wrapper over webkit. Give it a try if you have some time on your hands. It might not be convenient or simple in the beginning, but it can get quite interesting.

Case For C++

Abhijit Nadgouda — Fri, 29 May 2009 05:10:31 +0000

C++ has been at the receiving end in recent times about neither being a good low-level language nor being an effective high-level one. It is said that there is no case for C++ in today’s programming world. I agree that with evolution of programming languages, applicability of C++ is narrowing down, but it has still not vanished. Hypertable has a nice explanation of why C++ was chosen over Java. The rationale is specific to the context, but it is a good indication that C++ can offer a good combination of low-level control and high-level language features in memory and processor intensive applications.

The Whys Of Requirements Specification

Abhijit Nadgouda — Wed, 27 May 2009 14:22:16 +0000

One of the biggest gaps in the requirements discovery and specification, and solution design is formed because the whys are not communicated to the solution developers. The requirements specification usually talks only about what the solution should do. There is nothing that tells them why it is so. When their decisions are not aware of the underlying reasons, they can work against the overall benefit and at times make the solution a failure. Here are some things that I think should be communicated to the solution developers:

problems and pain points in the current system
priorities of requirements
business constraints
other non-software components, like processes and skills
rationale behind important decisions

A lot of businesses do feasibility studies but they are not shared with the developers. Or sometimes businesses have already identified a solution that they are using to build their requirements. Or sometimes businesses just do not invest enough in identifying their problems and pain points.

Unless the solution developers know the ROI behind the code they are writing of tools they are integrating, they are not going to be able to effect it. It is always better to define the problem in its own space, keep it solution agnostic and share it with people who are developing your systems. Or this gap can lead you to the wrong solution or a solution for the wrong problem.

Universal IE6 CSS

Abhijit Nadgouda — Fri, 22 May 2009 02:57:02 +0000

Andy Clarke has a new solution for dealing with IE6 incompatibilities - Universal IE 6 CSS. The idea is to serve simple design with great typography for IE 6, without layout. The Web developer community is trying various approaches to get rid of IE6. Andy focuses on serving readable content to an IE6 user instead of spending effort on trying to get everything to work in IE6 or completely blocking them away. I think its a great solution for public content sites, something that will save a lot of effort and still serve the IE6 user. I think Andy’s approach will face a lot of resistance in the corporate world, where IE6 is more prevalent and cannot be treated as a second-class citizen. But this might turn out to be the most efficient solution for public content sites.