Remarks On Twitter, the "Error 411" Issue, and Twitter Clients For Blind Users
Doug Lee
October 27, 2012
Last revised October 28, 2012, 9:15 EDT

Current Status

Sometime Sunday morning (Eastern time), twitter internally fixed the problem that upset several clients. The entire duration of the upset was roughly 36 hours. I regard Twitter's response as much slower than desired, yet I greatly appreciate the fact that the problem was fixed globally by Twitter so clients that did not manage to work around it can work again.

Claims and Comments

Here I collect and analyze a few assertions made on Twitter in the first 24 hours or so relating to the recent sudden failure of The Qube and other Twitter clients. Most clients that failed of which I am aware, though not all, are clients developed specifically for blind users.

I reserve the right to modify this document if/when I obtain new information affecting my view of its accuracy. If I make changes, the date/time stamp in the header will also change, and the nature of changes will be listed at the end of the document..

Claim

The Qube was incompatible with Twitter. Twitter didn't cause this problem.

Remarks

Twitter did make an API change on Friday, October 26, that was neither documented nor announced in advance as far as I can tell. Therefore I assert that this claim is false, and that Twitter did indeed cause this problem. It also affected a few non-blind-product-related pieces of software, as can be found from a Twitter search on "error 411" (without the quotes). I welcome evidence from anyone supporting the idea that this change was any of the following, and if any one or more of these hold true, I will likely rewrite these remarks:

1. The change was announced in advance, with approximate date or date range during which it would take effect, OR
2. The change was merely a new enforcement of a long-stated API requirement that Qwitter and its descendants (and other non-blindness-related clients) never followed properly, OR
3. The change was not a change made by Twitter itself but by some other intervening entity, such as an Internet service provider.

I assert that, if none of those assertions are true, Twitter itself did us a direct disservice that can reasonably be blamed for Friday's failures. The fact that they are making known API changes to go into effect next year is neither related nor mittigating.

Claim

Twitter's upcoming API changes are going to break permanently all the clients written for blind people.

Remarks

The upcoming Twitter API changes are sweeping and will undoubtedly, for each of many clients for blind and sighted alike, either break them or force significant updates to be made to their code. From Twitter's own documentation of their plans, it also seems likely that such major changes and forced updates will be a continuing trend for some time. Which clients will die and which will survive will depend individually on the developers of those clients. Developer responses to Twitter's changes will surely depend in turn on how the client audiences react to client failures, predicted shutdowns, etc. In my mind, there is no guarantee that any specific client, written for blind or sighted users, will survive indefinitely. We are dealing with probabilities here, and each of us must weigh what developers say they will do and what they have done against our own criteria for choosing a client. Personally though, I discourage people from choosing a client based on who it's written for, even in light of Twitter's expressed intentions to complicate continued development of clients that, as part of their function, aim to alter the Twitter interface. I recommend choosing a client based on your personal needs, the developers' stated goals and plans, and the track record shown by those developers on holding to those goals and plans. After all, using future likelihood of client functionality as a criterion really has nothing to do with blindness at all. Sighted people want that too.

Claim

Blind people are better off using a mainstream client and getting away from blind-specific clients like The Qube, Qwitter, and Twitmonger.

Remarks

The idea here is that mainstream clients will be better maintained because of the wider audience. The claim is a fair bet but is neither a guarantee nor well supported by current indications. See the next claim for more on the second part of that.

Claim

Tween is discontinued.

Remarks

The company formerly responsible for developing Tween, Ubermedia, announces the product's discontinuation prominently on its web site. (See http://tweenapp.org.) However, the current concensus is that the actual developer of Tween is continuing work on the client. (See https://sites.google.com/site/tweentwitterclient/.) I believe this is good but also that it reduces Tween to the same status as the so-called blind clients: It is maintained informally, without compensation, by one person or a small group of persons, and may disappear at any time, just like Qwitter, the Qube, etc. I therefore assert that Tween itself is arguably no longer a "mainstream" client. The developer does promise continued work on Tween, to be fair.

Claim

There has been about as much bitterness and invective as information posted on Twitter about this client issue.

Remarks

This claim is mine, made here on this page for the first time. I make it based on what I have seen in the last 24 hours or so on Twitter, and I confess profound disappointment in the blind community over it. The choice among clients is just that, a choice, no more and no less. As we have the right to choose computers, clothing outfits, and Presidential candidates for which to vote, we have the right to choose among clients. All choices have pros and cons. Those who choose are responsible for evaluating these as always. But my choosing Qwitter, or the Qube, or Tween, or Tweetlist, does not render me categorically a fool nor a sage. It renders me a man with a viewpoint, a decision, and hopefully, a means to talk on Twitter.

Further Reading

List of Changes Made Since This Document's Initial Publication

October 28, 09:15 EDT

Added the Current Status section and a Claims and Comments header above the main section, and adjusted language like "yesterday" slightly to be more specific now that more time has passed.

October 27, 17:50 EDT

Reference section added with two entries.

October 27, 14:30 EDT

Initial publication.