Does Spelling Matter for the Real-time Web?on Mar 12, 2010 in Real-time web by Ian Alderson
A few weeks ago a colleague sent through a link to this blog by Paddy Donnelly (warning, this link carries a parent advisory sticker) in which he expresses his frustration with web sites littered with spelling mistakes. I also find these mistakes annoying, although I am treading on thin ice as I am a frequent perpetrator of using effect instead of affect.
Within a few minutes of reading this blog, I sent a tweet that contained several spelling mistakes. With Paddy Donnelly’s blog still on my mind, this made me wonder whether spelling mistakes matter so much within the real-time web? Does the importance of sending out an update now outweigh that of ensuring it is spelt correctly and is grammatically correct?
For me, the purpose of the real-time web is to be informed of events “now”, as they happen. As a result I am happy for a few typos or spelling mistakes to creep in, rather than risk that data arriving late. However, there are a couple of sub-clauses:
- Mistakes must not change the meaning of the information, making it either incorrect or misleading
- Mistakes must not render the information unintelligible.
Example: Traffic Updates
A real world example of this is traffic updates. A number of years ago, before a mobile real-time web was possible, I was listening to the radio to obtain traffic updates whilst driving on the M25. These updates were coming in every 15 minutes, which back then seemed perfectly acceptable.
Unfortunately a granularity of 15 minutes turned out to be inconvenient as the traffic report that came in to say that there were heavy queues due to an accident between junctions 8 and 9 clockwise was received at exactly the point I was driving past junction 8. Now it could be that the accident had only just happened and this was genuinely the earliest I could have been informed of it. However if it had occurred a few minutes earlier then I would have been able to come off at junction 8 had I been warned about it sooner.
The Real-time Web
With the real-time web I want information to be available to me immediately; I don’t want it encumbered by several seconds, or more, of spell and grammar checking. The relevance of the information may well decay over time, like some kind of data half life.
The sources of real-time information are also a potential burden. It seems to me that a decent proportion of the real-time web snippets of information are sent from mobile devices which, in my personal experience, are a lot more prone to typos and spelling mistakes.
Correct and Valuable Data
As mentioned before an important caveat for the importance of now is that the data must be correct. In my example of the traffic report a simple mistake where the M25 is misreported as the A25 or M26, or where the junction numbers are incorrect, renders the information useless to me even if I receive it immediately.
I believe that there is a happy medium between spelling mistakes and immediacy. When something has happened we should be finding out about it now. Once this has been addressed the spelling mistakes and grammatical errors can be corrected and the information embellished. All of this should seem familiar since it’s effectively what news agencies have been doing for many years.
Financial Data – a working example?
This type of behaviour is already something that I am familiar with in the financial software industry. Various data providers, such as Thomson Reuters, take data from many different sources (e.g. Stock Exchanges) and clean it up, removing outliers and obvious mistakes before sending it out to their customers. Some financial institutions, however, perceive there to be an advantage to taking direct feeds from the source instead so that they avoid losing valuable fractions of a second on the data being published. They are willing to have to deal with the “dirty” data because it is less costly than the time lost in order to receive “clean” data. The same could be said for the real-time web.
When is now?
My concise answer is as soon as reasonably possible. It partly depends on the frequency of updates to the state of something, and what the impact of not finding the information out soon enough. For example, if I am interested in finding out when my favourite author has published a new book, finding out to a precision of an hour, or even a day or week, might be appropriate. However if I want to know what the latest score is in a sports match, then I might only be willing to wait a few seconds.
For any data where I am only willing to wait a few seconds for, the only viable technical solution is for the data to be pushed to the user.
For me the immediacy and accuracy of information outweighs the need for a message to be free from spelling or grammatical mistakes. These are still important, however they should not impact the ability to send your message now. Once things have calmed down it’s time to go back and correct those mistakes, since the message is no longer about now, it’s about then.