If a service launches based on another service that's always broken, does it make a sound?

A message came across my Twitter feed this morning, posted via TwitterFone.  This service allows you to call a local number from your cell phone and record a short message, which is then auto-transcribed into text and posted to your Twitter feed.  Don’t get too excited, it’s invitation only right now.  (Here’s a sample message)

As a million and one different twitter-based services pop up, Darwinian forces will pull some into mass circulation as others become relics you can eventually laugh at your friends for inviting you to join.  So far twitpic and twitterberry seem to fill acutal niches (autoposting photos and twitter access from your BlackBerry, respectively) and - all knowledge gathered via my own anecdotal spidersense - these seem like they’re here to stay.  Tweets sent from desktop apps like Twitteriffic also seem to have some fans, but most standalone software appeals to the more niche market of those that want to cross post items from another RSS feed (twitterfeed) or Pownce (Twhirl).  Like any evolutionary process, my guess is that the more specific the software, the more likely it is to die out as the environment changes.

We all recognize the uses for Twitter itself - for entertainment, microblogging, networking, pushing information, etc etc etc… - but we have also all encountered that damn whale (apparently a stock photo) and vast spans of downtime.  So how can any service, especially one that relies on a technology as imperfect as voice recognition and gives up the super-fast, super-discreet traits of its parent, expect to develop a large and loyal userbase?  Getting a tweet to post in its entirety is important to many people; do you know when you’ve spoken 140 characters?  It’s like multiplying fractions…. if Twitter is up 95% of the time (random generous number) and you send 10% of your tweets as voice messages, and 50% of those are translated correctly, and half of those have enough in the first 140 characters to generate a click-thru… does it really matter?  If you usually attempt to tweet a message every 30 minutes you’re awake (let’s say 18 hours), that still works out to .855 successful TwitterFone tweets a day - in otherwords, maybe a single tweet hits its target.  Is that enough to perpetuate its existence?

If you have an opinion, tweet it out.