Developing applications can be interesting, entertaining and challenging - you never stop learning. Here's an example of how new features can cause new problems that can be solved with new features and so forth...

I wanted shortlinks

Often links are too long to send with microblogging services. If a user wants to publish a link, he has to use a shortlink then. For twitter users, services like tinyurl or twitturly are popular for this task. Some time ago, I added the shortlink feature to zzap. It can be used in two different ways.

The easy way is to click on 'attach shortened link' under the message field. A new input field for the (long original) link you want to publish appears. After the user clicks on the 'send' button, a shortened link of 31 characters is included in his original message.

The not so easy, but for me more useful way is to use the shortlinks for the announcement of blog entries. I'm using a ping from my blog to announce the new entry to zzap. zzap determines the (long!) original link of the blogentry, shortens it and posts a message with the headline of the article and the shortlink.

What were the reasons for shortlinks?

The reasons, why I wanted to integrate such a feature in zzap were:

• More convenience for the zzap user
• Independence from third-party services like tinyurl
• Better information for the zzap user

The last was the most important reason for me. I wanted to know, how many of my followers were interested in my blog articles, what kind of devices they are using and in what timespan my posts are 'hot'. Soon I gained some basic informations.

• A high percentage of my follwers were clicking on the shortlinks - far more than I expected!
• Headlines in english were most popular - although most of my followers are german.
• During the week, most clicks came imediately within the first minutes. At the weekend they came within two days.
• Most users are viewing my messages with a normal browser.
• However - some were using mobile phones, or the iPhone

Shortlinks causing new challenges

The last information demonstrated the growing importance of mobile internet usage. But it also showed a weakness of the shortlink features. If one clicks on the shortlink on his mobile phone he will be redirected to the original URL. And this page is normally built for normal webbrowsers. It's no fun to view such pages on a mobile. Yes, I know - the iPhone can display them...

A new feature - again...

While I can't do much for external links, I wasn't satisfied, that my own blog anouncements aren't useful for mobile users. I thought about, what I could do to make my own blog "mobile device enabled". The good news is, that I found a solution for my favored blogging system serendipity. Now I'm testing the ADD (Automatic Device Detection) and ACA (Automatic Content Adjustment) with this very blog. When no problems occure during normal operations, I will update my personal blog too and write a detailed manual on how to upgrade serendipity for mobile reading.

From the very beginning, zzap supported mobile devices. Even the prototypes from 2006 and 2007 did. Since May 2008, there's not only a version for mobile phones, but also a special iPhone version.

But in the last few days, asked myself, if it is really necessary. Sure - it's cool, but not trivial to maintain. This is, because coding with the iui library is very different than the ordinary XHTML Pages for the web-version and the mobile-version. The old WAP-days came back to my mind, with the 'deck-paradigma'. There's also a lot of Ajax magic in it and the form processing is - uhm - special. This all causes not only a lot of view-templates, but also changes in the contoller logic. A lot of additional work at the end of the day and I'm still the only developer in this project with very little time.

Maybe the normal pages for mobile phones are good enough? The functionality is the same. Maybe I could polish it with a special CSS for the iPhone. What do you think about it?

The last few weeks were a bit busy for me, therefore the progress on zzap is rather slow at the moment. But nontheless yesterday I published version 0.4.142. Not a big jump but there are some improvements. The visible part for the user:

• When answering to a message from a twitter account, the '@username ' is automatically inserted
• A simple reporting for the shortlinks is inluded. One can see, how many people clicked on the hortlink, how many were guests or members, and what kind of device they were using (browser, mobile, iPhone...)

Of course, there are some improvements under the hood too, mainly concerning the blog plugin. It is now nearly stable and only one function is missing: sending zzap-answers to an announced blogentry as a trackback. That seems to be a tricky part as Wordpress and Serendipity are a bit fussy.

There are still so many issues on my list... **phew**

Users are funny - few features are too few, but many features are to many. I slowly understand, how microsoft office became such are bloatware.

More?
During my study in 2006/2007 I developed a Software for short messaging via web and mobile phone. I found, that most people are using their mobiles for the following reasons: Chat, making appointments and changing appointments. Therefore I thought features to support the following issues would be useful:

• short messages (You'll keep short while you're on your way.)
• group functions(One message for all of your friends.)
• location functions (Where's that cool party?)
• schedule (when will we meet?)
• Information about your situation (Look - it's so nice here...)

I had built a prototype, but it turns out, that it's too complicated. None of my test users used one the 'higher' features.

Or less?
Then came twitter with an extremely limited 140 character messages service. It became popular, just because it was so focused, with only one main feature.

Or what?
After a while, the users got bored and soon new features around twitter appeared: Photo upload (twitpic), post links (tinyurl) and a few days ago I found a service for posting music files. Now Don Reisinger demanded on TechChrunch ("Why Twitter Needs to Do More") group functions. Great - then the actual state is similar to what I had one and a half year ago, only more complicated.

What's the lesson to learn from that?

When laconi.ca - the software behind the microblogging service identi.ca – was released, I was working on the current (4th) incarnation of zzap. When I compared the two products, I soon realized, that our goal was similar, but our approach was different. I guess it's because we've weighted several features using different importance.

Still missing: a microblogging exchange protocoll
For example: Laconi.ca started with a protocol for the communication between individual instances of the software. I had a similar idea - on my papers (yes – real papers!) were sketches, XML- and method definitions of a lightweight protocol for remote subscriptions and so forth. Inspired by the planned subscription process, I chose the working name 'pingpong'. This protocol was important, but other features were even more important for me. I asked myself „how can this software be useful for me as soon as possible?“

From the very beginning: mobile support
To me, microblogging is only useful, if it supports multiple clients in different enviroments and situations. I want to use it spontaneously, while I'm on my way. That's the reason why one of my initial goals was the support of mobile phones. My own framework has an integrated device-detection, and supports normal web, mobile web and (to satisfy my own curiosity) the iPhone/iPod touch.

Visual impressions: Photo support
In certain situations, while I was on my way, I wanted not only to send a tiny message from my mobile phone, but also attach a photo, to show why this situation was special or funny. To statisfy my own demand, I included a photo upload field and a mechanism to fit the dimensions of the photo to the viewing device. Mobile phones with a tiny display reveive smaller pictures, than phones with a bigger display. The biggest pictures are for the web version.

The existing network
I had some followers on twitter and was tired of using zzap and twitter one after another. It would be better to push my zzaping to my twitter account and pull the new messages from the people I follow on twitter. A twitter integration was therefore more important than 'pingpong'.

Spread the news
As I realized, that laconi.ca had already solved the problem of remote subscriptions, I decided, not to re-invent the wheel and adopt their protocol later, considering that others had also announced to do so. It's better to support an upcoming standard. That in mind, I focused on a different demand: A better integration between my blog and my microblog. That's why I started to build a function to announce new entries from my blog with an automatic included shortlink like tinyurl.

There are a lot more ideas to be implemented. My main goal is to support an upcoming microblogging standard and build the features, that I and my friends want. And by the way; considering how we all use twitter & co, microblogging isn't the right description. We deal rather with a kind of group-messaging. Just my point of view.

I just updated www.zzap.de to version 0.4.127. A short explanation of this Number: The leading zero has the meaning: not stable for a production enviroment. The four in the middle show, there were three previous versions of zzap (that I don't want to write about now - maybe later). The last 127 is the revision number in the Subversion Repository.

What's new?

• You can switch between English and German on every page. If you're logged in, this switch overrides, but doesn't change the settings in your Profile.
• You can switch between the web version and the mobile version on every page. There was a demand for that feature, because in some mobile browsers (e.g. opera) it can useful.
• The most powerfull new feature lies under the hood: A brand new plugin system for external communication.

While the first two changes are rather small, but useful for the convenience of the users, the third is the key for further development. One main aim for zzap is the integration into third-party services. To ensure a flexible, quick and easy way to integrate new services for outgoing and incoming messages, a kind of plugin system was neccessary. I did the rewrite mainly last week and last weekend. Until now, there are two plugins: one for twitter and one for the blog announcements. More are planned for the future.