seah@ee.rochester.edu (David Seah) (07/17/90)
I'm interested in finding an article or book that compares the relative virtues of using a noun-verb command structure (ala the Apple Finder) versus a verb-noun structure (like workstation-based CAD programs I've come across). I've been searching around Computer-Human Interaction type subjects, but I suspect I should start looking into psychology more. Any pointers? --- One of the benefits of a noun-verb or object-action command structure, according to the Apple Human Interface Guidelines, is that the "paradigm matches the syntax that we normally use in ordinary noncomputer actions: 'Hey, you...' (selection) '...do this' (choose an action)." Is this really the case? I tend to think that this syntax comes with a certain way of thinking about the computer system in general. With the noun-verb paradigm, the computer is treated as an unintelligent box. You must alert the computer, by nature a dull and listless creature, by playing with the objects on the screen to attact its attention (selection). You then command the bitbox to do something with those objects (choose an action). The "ordinary noncomputer" syntax seems suited for playing with the dog than with other people. The verb-noun paradigm seems to be more geared toward the notion of an intelligent computer. It's certainly more interactive, where the computer is on a footing equal to yourself (sort of :) You say, "Arrange those boxes!"(choose an action). The computer responds, "which boxes?" (selection). Then it does it. Or perhaps it doesn't, which is more realistic. Is the idea of a subordinate computer the idea behind the noun-verb paradigm? Personally, I like noun-verb better because it doesn't commit the user toward performing an action, like not letting go of a chess piece until you're really decided on your course of action. It's also a little like my RPN HP calculator in that it saves "user-strokes". Verb-noun command structures tend to wallow in obscure modes, and are tedious to use because they are more suited toward the batch-processing mentality. Sorry if these topics were covered in Basic Cog-Sci 101...I'm a EE that is just getting interested in these topics. -- Dave Seah | Omnidyne Systems-M | INET: seah@ee.rochester.edu ^..^ | "User-Friendly Killing Machines" | America Online: AFC DaveS yargh! +----------------------------------------------------------------=*
ogden@nmsu.edu (Bill Ogden) (07/19/90)
Jackson (1983) compaired performance of users of verb-noun commands (e.g "Find the VW ads") to users of noun-verb commands (e.g. "VW ads find") and found that despite the naturalness of the verb-noun syntax, there was very litte difference between the two groups. This was true for both experienced and novice users. This suggests that the experience using natural English does not transfer to learning and using a computer command language. It also suggests that the experience of using a verb-noun computer languages (which matched the UNIX experience of the experienced group) does not negatively transfer to learning and using a noun-verb language. Jackson further found that users often left out important parts of their commands (e.g "VW find" vs "VW ads find"), suggesting that semantic restrictions are harder for people to learn than syntactic ones. Reference: Jackson, M. (1983) Constrained languages need not constrain person/computer interaction. SIGCHI Bulletin, 15(2-3), pp. 18-22. Bill Ogden Computing Research Lab New Mexico State U. ogden@nmsu.edu
warner@scubed.SCUBED.COM (Ken Warner) (07/19/90)
>Jackson (1983) compaired performance of users of verb-noun >commands (e.g "Find the VW ads") to users of noun-verb >commands (e.g. "VW ads find") etc..... >Bill Ogden Noun-verb looks a lot like message passing in Smalltalk. In particular, unary messages (method selectors of methods without parameters) look very much like the example. VW would be the object; ads would be a method that perhaps would return a collection of ads relating to VW's. The difference, not illuminated in the example, would be that VW might be an element of a set of cars and find would be redundant as the method "ads" would return the ads. So in Smalltalk: Cars VW ads. would be the statement one would use. After a couple of years of Smalltalk'ing, I find noun-verb at least as evocative as "natural language." Maybe I should go out more. Ken Warner
scottl@mercury.sybase.com (Scott Luebking) (07/30/90)
Another way to look at the noun-verb order concerns the number of nouns which are involved. I think there are three cases which are: 1. one verb and no nouns 2. one verb and exactly 1 noun 3. one verb and 0 or more nouns The first one is the trivial case where the verb doesn't operate on any nouns. The second case is one where a noun must be provided for the verb. In this case, both pieces of information are needed before any error checking is done or any resources, e.g. locks, printers, are requested. Since both pieces are needed before any processing can continue, the order may not be as important. One consideration might be that if the verb is chosen first, the interface could indicate appropriate noun choices, e.g. highlighting or popping up menus or text fields. Similarly, choosing the noun first could allow the system to indicate appropriate verbs. The third case is used to specify several nouns. It is useful for specifying a list of objects where the action is performed on each object, e.g. printing a file. For an expert user, it increases effificiency because he/she specifies the verb only once instead of once for each noun. A list of nouns can alo be used to specify a relationship among the list of nouns or build a noun aggregate, e.g. specifying a group of the objects. Specifying a list of nouns can have one of the two following forms: <noun> <noun> ... <verb> <verb> <noun> <noun> ... <terminator> In the first form, the verb has the dual purpose of specifying the action and terminating the list of nouns. In the other form, a terminator is used to signal the end of the list of nouns. The main advantage of the first form is that the user doesn't have to specify a terminator. The second form has several advantages. One advantage to the second form is that error and resource checking can be performed as each noun is selected rather than after the entire list is specified. Checking each noun as it is selected is much less frustrating for the user since the user knows immediately that there is a problem rather than at after specifying the entire list. If the error or resource checking occurs at the end of the list, the user will need to specify the entire list again. Also, if more than one noun has an error, pairing errors and nouns after the list is specified can be confusing or time consuming. Another advantage to the second form is that more support can be provided to the user when specifying complex syntax. Complex syntax has the advantage of improving user efficiency by increasing the brevity of commands. In this example: save noun-1 to noun-100 except noun-34 the "to" and "except" are nouns used to modify and shorten the command. The syntax uses fewer nouns than: save noun-1 noun-2 ... noun-33 noun-35 ... noun-100 Complex syntax also increases the expressability of the interface. In more complex syntax, the relations between the nouns and the order of the nouns becomes much more significant. The example could be written as: noun-1 to noun-100 except noun-34 save However, the interface cannot prompt as easily because it doesn't know the verb. Since the interface cannot assist the user through the complex syntax, the user will be forced to learn and remember the syntax. This user requirement will increase the amount of effort the user will need to exert in order to take advantage of the complex syntax. The result will be that few users will take advantage of the complex syntax which will thwart the purpose of the syntax. If the verb is at the beginning of the noun list, the interface can assist the user through complex syntax by prompting. The second form is much more user-friendly and can provide facilities which can significantly help expert users. If the application is simple with no noun aggregates and need not worry about expert users, the choice between noun-verb or verb-noun order can be made on other criteria. If the application is complicated or will have expert users, the verb-noun order is probably preferred because it can be expanded into supporting noun aggregates or more complicated syntax. -- Scott Luebking {mtxinu,sun,pyramid,pacbell}!sybase!scottl scottl@sybase.com
brucec@phoebus.phoebus.labs.tek.com (Bruce Cohen;;50-662;LP=A;) (07/31/90)
There is one advantage to non-verb for _sequences_ of commands on a given object: you don't have to constantly rename the target object. So a sequence might be noun-verb-verb-verb. The utility (or even desirability) of this obviously varies with the application. I use sequences like this often when using a drawing program to make complex, multi-object-group diagrams. On the other hand, the inverse kind of sequence is also common: perform the same operation on multiple objects. Many Apple-style drawing programs do this by having you select an operation from a menu of operations, and applying that operation to any object you point at until you select a different operation. I suspect that this choice is somewhat less driven by abstract CHI principles than by the dictate that the Macintosh had to use a one-button mouse. It seems clear to me that both styles should be available at any time to a user, since the choice is dictated more by the immediate work flow than by the kind of work (which application is being used) or by "goodness" of UI design. -- --------------------------------------------------------------------------- Bruce Cohen, Computer Research Lab email: brucec@tekcrl.labs.tek.com Tektronix Laboratories, Tektronix, Inc. phone: (503)627-5241 M/S 50-662, P.O. Box 500, Beaverton, OR 97077
ianr@syma.sussex.ac.uk (Ian Rogers) (10/19/90)
I seem to remember this thread from before, but I've managed to lose the conclusions. Briefly the question was: is there any effective (or even affective for the philosophers out there) difference between the Noun-Verb and Verb-Noun interface metaphors. [The Mac is Noun-Verb: eg. you pick up a file (Noun) and put it in the waste basket (delete, Verb), whereas the (at least) graphics part of FrameMaker is Verb-Noun: you select line drawing mode and then draw one. Command line languages are Verb-Noun] I'd love some references (preferable) or educated opinions (also valuable). Email, and I'll post if there's interest. Thanks in advance, Ian Rogers Janet/Arpa: ianr@uk.ac.susx.cogs | Cognitive & Computing Sciences uucp: ukc!cogs!ianr | Sussex University, Falmer, voice: +44-(0)273-606755 x2392 | Brighton, England. "This has been an ACME Straight To The Point production" - me
dmark@acsu.buffalo.edu (David Mark) (10/20/90)
In article <3648@syma.sussex.ac.uk> ianr@syma.sussex.ac.uk (Ian Rogers) writes: > >Briefly the question was: is there any effective (or even affective >for the philosophers out there) difference between the Noun-Verb and >Verb-Noun interface metaphors. > > [The Mac is Noun-Verb: eg. you pick up a file (Noun) and put > it in the waste basket (delete, Verb), whereas the (at least) > graphics part of FrameMaker is Verb-Noun: you select line drawing > mode and then draw one. Command line languages are Verb-Noun] > >I'd love some references (preferable) or educated opinions (also valuable). > Noi Sukaviriya and Lucy Moran, grad students in Jim Foley's group at George Washington University, conducted a very interesting study of verb-object or object-verb PREFERENCES among subjects. The study was published in: Sukaviriya, Pinyawadee, and Moran, Lucy, 1990. User interfaces for Asian countries. In Nielson, Jakob, editor. Designing Interfaces for International Use. Elsevier Science Publishers. Briefly (and from memory), they tested subjects who were native speakers of English, of Thai, and of languages from the Indian sub-continent. (Like English, Thai expresses most commands and queries with a verb-object sequence, whereas the Indian language(s) tested were the opposite.) Subjects were not told what order to use in a drawing/coloring task. Some were tested with direct manipulation (mouse) and others with keybord entry. There were no significant differences between Thai and English speakers' sequence preferences, but the Indian-language speakers reveresed the prefered sequence on at least some tasks. I should probably stop here, and refer you to the authors, or their article, or find my copy! But the point is that user preferences seems to be associated with word-order for the users' native language and with the exact task. This seems to be a very rich and hardly-explored topic within HCI/CHI! David Mark dmark@sun.acsu.buffalo.edu