I have just finished Peter Bøgh Andersen's "A Semiotic Approach to Programming". Andersen uses a very grammar-oriented, structuralist Hjelmslevian flavour of semiotics. The essay seems to be rigid and thorough to me, hobby-programmer as I am, using a selection of Hjelmslevian concepts to launch a programming style.
It seems as a side-track for semiotics, however. To me, semiotics is first and foremost about the conditions and creations of meanings, not computer programs. And Andersen's concept of sign is weird. A sign can be a whole computer system, a single letter, or anything in between. When Andersen understands Peirce's concept of interpretant as "the speech and actions of the user in response to the execution [of a program]", he is sweeping the whole issue of human understanding, of meaning itself, under the carpet, reducing the semiotic approach to a reminder that computer code is something humans read from time to time.
Moreover, Andersen asserts that computer signs (on the screen this time, and not code, I guess) are able to change over time. His favourite example is the grayed-out buttons in Mac-style GUIs. To me, a greyed out button is a different sign than an active button. The active button signifies that a process is available, a greyed-out button signifies that a process is not, but may be at some other point.
To Saussure, Peirce, Eco, and many others, signs are relations between signifiers and signifieds (and, to Peirce and Eco, the difficult interpretants). If the signified changes, the whole sign changes. Even if both resemble a button.
<< Previously in Surftrail:
Über-box
Next: >>
Fun Factor