laser-lovers@uw-beaver (05/18/85)
From: Les Earnest <LES@SU-AI.ARPA> As previously advertised, the Metafont Lunch Bunch had a blind tasting of two fonts today. There was a fairly strong concensus preferring one of them, by 11 to 3, but several participants expressed a desire afterward that the winner and loser not be identified. I will honor that request. There was also fairly universal agreement that this was an interesting experiment and that it should be repeated with a wider selection of samples of the (nominally) same typeface, perhaps including several sizes, representing as many different font source/engine combinations as possible. During the next week we will formulate some specifications and source-language text samples, which will then be dispatched to Laser-Lovers in the form of a solicitation of print samples. In case some other groups wish to try their own tastings, I will outline the procedure that we used. Incidentally, 16 people showed up today. Before the tasting began, the group discussed the question of what criteria should be used in ranking the samples. One possibility was relative "readability." Another was "authenticity" (i.e. relative closeness to the "ideal" Times Roman). The group finally settled on nonspecific "preference." In other words, each individual was asked to choose between the samples on the basis of whatever criteria he/she thought important. In order to avoid one possible source of bias, any participants who concluded that they could identify the source of any sample were asked not to vote. As the supplier of samples, I was of course excluded. One other participant also disqualified himself. To start things off, we asked a person who did not know the origins of the samples to mark arbitrary identifiers on each and place them in a manila folder that also held a small tablet to be used for ballots. Don Knuth took on this responsibility. I later discovered that he had gotten into the spirit of winetasting by marking one sample "red" and the other "white". The folder was sequentially passed among the participants, each of whom marked their preference on a separate "ballot", which they folded and passed to the designated clerk. As administrator, I got to see the samples last and announced which was which. (In order to keep the administrator honest, anyone who wished could take a copy of the print samples for comparison with other sources.) Finally, the ballots were counted and the results announced by the clerk. That's the scheme we used. We may refine it a bit for our next tasting, which will happen after we have solicited and received a wider selection of print samples. Cheers, Les Earnest