Tuesday, December 15, 2009
Infrequent Updates
I'm in the midst of a rather demanding project for an education company, so trying to keep to a weekly schedule is going to be difficult through to the end of January. I'll add things as I can and as the mood suits me.
Friday, December 4, 2009
Feature Smell
My whole point in this blog is that I have this vague feeling that computer language instruction tools are not filling all their potential. My only background in this is as a software designer who has studied a number of foreign languages. But in using software to learn a language, I'm finding many aspects of the software that leave me wondering if the design is entirely thought through. In effect, I'm experiencing a sense of Feature Smell.
In software design, coders speak of Code Smell to describe aspects of the written software that suggest that something needs to be redesigned. Nearly-identical repeated lines of code or a very long method are examples of code smells. In general, a code smell needs to be cleaned up, though there might occasionally be a good reason for it to exist.
As a software professional, I tend to look at any piece of software I use and wonder what the designers and implementers were thinking when they created it. As a result, I've begun to wonder whether there are feature smells that point to larger design problems. These would be identifiable features that suggest places where the design needs to be reviewed. If so, I think I see some in standard language learning software.
Limited and Unchangeable Content
Most language learning software I've seen provides a set collection of words and phrases they teach to the student. The phrases you're given can't be enhanced or altered and usually provide a very limited vocabulary in the end.
In most software, a limited and unchanging content is usually a sign that too much of the design was done with only one purpose in mind, and that every attempt to expand the software was done as a one-off implementation. In this case though, I think that the problem has more to do with the original model of the learning tool.
Language learning software was originally sold as an extension of language tapes. Software was bought as a standalone product that had no connection to the web. It was not designed to be updated over time. The content is limited because it is carefully crafted, voice actors have to be hired to speak the words, and adding new content requires the student to buy more media. The result is that sample phrases from Rosetta Stone such as
In an actual classroom, the teacher will use a standard text and supplement it with other content, such as articles, stories, children's books, videos, radio programs, and so on. So far, I have seen no software step away from the language tape model and provide links to this sort of additional content.
There are other odd features, particularly in the way drills and tests are handled, but I think I need to talk to a few more teachers before I decide whether that's a real problem or just a pedagogical technique I don't yet understand.
In software design, coders speak of Code Smell to describe aspects of the written software that suggest that something needs to be redesigned. Nearly-identical repeated lines of code or a very long method are examples of code smells. In general, a code smell needs to be cleaned up, though there might occasionally be a good reason for it to exist.
As a software professional, I tend to look at any piece of software I use and wonder what the designers and implementers were thinking when they created it. As a result, I've begun to wonder whether there are feature smells that point to larger design problems. These would be identifiable features that suggest places where the design needs to be reviewed. If so, I think I see some in standard language learning software.
Limited and Unchangeable Content
Most language learning software I've seen provides a set collection of words and phrases they teach to the student. The phrases you're given can't be enhanced or altered and usually provide a very limited vocabulary in the end.
In most software, a limited and unchanging content is usually a sign that too much of the design was done with only one purpose in mind, and that every attempt to expand the software was done as a one-off implementation. In this case though, I think that the problem has more to do with the original model of the learning tool.
Language learning software was originally sold as an extension of language tapes. Software was bought as a standalone product that had no connection to the web. It was not designed to be updated over time. The content is limited because it is carefully crafted, voice actors have to be hired to speak the words, and adding new content requires the student to buy more media. The result is that sample phrases from Rosetta Stone such as
The boys are running. They are not walking.has replaced
La plume de ma tanteor
My tailor is rich.as an example of a phrase that everyone ends up knowing, but will probably never use in real life. Even an online service like LiveMocha has this problem. There, content is often directly translated from lessons for other languages, leading to oddities like a preference for the progressive tense in languages that use it far less than English.
In an actual classroom, the teacher will use a standard text and supplement it with other content, such as articles, stories, children's books, videos, radio programs, and so on. So far, I have seen no software step away from the language tape model and provide links to this sort of additional content.
There are other odd features, particularly in the way drills and tests are handled, but I think I need to talk to a few more teachers before I decide whether that's a real problem or just a pedagogical technique I don't yet understand.
Subscribe to:
Posts (Atom)
