Order Toll Free 1-877-339-5855
News, How-tos, and assorted Views on Accordance Bible Software.

Friday, July 13, 2007  

Lexical Forms Controversy!

On Wednesday, I wrote about the distinction between searching lexical forms and inflected forms in Greek and Hebrew tagged texts. In response to that post, an anonymous commenter wrote that this is an example of bad interface design, primarily because users need to know whether or not they are searching a tagged or untagged text, and must adjust their searches accordingly. This user argued that a better design would be to make inflected forms the default search behavior, and then to require the user to do something explicit to search for lexical forms. This, he argues, would eliminate inconsistency and confusion when using the same search on both tagged and untagged texts.

It's a fair criticism, so let's consider it a moment. Let's set the issue of working with multiple texts aside for now, and consider what most users typically expect when working with a single text. Don't worry, we'll get to the multiple text issue in a moment.

Why do we make searching lexical forms the default behavior, and require you to enclose inflected forms in quotation marks? Because users are more likely to look for every occurrence of a word no matter how it is inflected than they are to look for a particular inflection. Let's say I'm a first-year Greek student and I've just learned that John 21 uses two different words for "love": agapao and phileo. I then fire up Accordance, open my GNT-T, enter agapao, and click OK. Since Accordance defaults to searching lexical forms, I get what I am looking for: all occurrences of agapao no matter how they are inflected. Were inflected forms the default, I would get an error message telling me the inflected form agapao does not exist in the Greek New Testament. "What do you mean agapao is not in the Greek New Testament?," I yell at the computer, "My Greek professor just spent a half-hour talking about it!"

The point of this hypothetical scenario is simply this: if you're more likely to want to do a lexical search rather than an inflected search, good interface design demands that you be able to do it without having to learn some complicated search syntax. Not only would such syntax confuse the new user, it would burden the experienced user with the hassle of having to do something extra to accomplish the kind of search he performs most often. Obviously, I use Accordance all the time, and I can't remember the last time I wanted to search for a specific inflected form. I certainly don't want that to be the default behavior.

Now, let's go back to our search for agapao and consider what happens when you switch to an untagged text. The user who raised this issue mentioned the NA27-GBS text, which is an untagged version of the Greek New Testament included on the German Bible Society's Mac Studienbibel CD-ROM. This package, which is published by GBS, contains the Greek and Hebrew critical apparatuses, along with untagged versions of the Greek and Hebrew texts containing the sigla (the little footnote symbols) which point to the apparatus. Thus, even though the NA27-GBS is untagged, it is useful to have displayed so that you can see the sigla marking the existence of variant readings.

Okay, so let's say I switch my search text from the GNT-T to the NA27-GBS with agapao still entered in the argument entry box. When I click OK to perform the search, I get an error message telling me agapao can't be found. That's because the NA27-GBS is untagged and so can only be searched by inflected form. As I understand it, this is the basis of our anonymous commenter's complaint. He is trying to use the same search on the same text (the Greek New Testament), but he gets a different result. The point that needs to be understood here is that not everything is the same: the NA27-GBS is a different module which lacks all the information contained in the GNT-T, so is it bad interface design that searching these different modules returns different results?

The simple solution to this alleged "inconsistency" is to avoid searching the more limited untagged text. If this user has both texts, he can confine his searches to the tagged GNT-T, and simply display the NA27-GBS in a parallel pane so that he can see the sigla. Like this:

I can see no reason why one might want to search the NA27-GBS separately from the GNT-T, but if you really want the NA27-GBS to be displayed in a separate window or tab and you want to see the search hits highlighted in the text of the NA27-GBS itself, there is a way to do it: use the HITS command to display the results of GNT-T searches in the text of the NA27-GBS.

In the screenshot above, I have a tab containing my search for agapao in the GNT-T. I then have a second tab containing the NA27-GBS. In that tab, I have entered the HITS command to join the two tabs. That way, the results of every search I do in the GNT-T are displayed in the NA27-GBS! The HITS command lets me leverage the tagging of the GNT-T to search the untagged NA27-GBS.

That, of course, is a fairly sophisticated set-up, and hardly what most users are going to want. But the capability is there for power users who want to search untagged texts.

After twelve years of helping to develop Accordance, I've learned that good interface design, like all forms of engineering, is a balancing act. Ultimately, you have to make choices based on what most users will do most of the time. Inevitably, some users will find those choices frustrating or confusing—particularly if they are used to another way of doing things. In many cases, user's complaints and suggestions can help us to see places where our interface choices have been wrong—instances where most users do things we didn't necessarily anticipate. In other cases, we remain convinced that we've made the right choice and we ask the user to adjust to those choices. Even then, where we are able, we look for ways to make that adjustment as painless as possible. In my next post, I'll show you an easy way to avoid the confusion which arises when a user copies an inflected form from the text, pastes it into the argument entry box, and then gets an error message because Accordance defaults to searching lexical rather than inflected forms.

It may or may not be true that "first year Greek students" want to search on lexical forms (are first year Greek students the target audience?), but I don't think it is true that users new to a bible program EXPECT it to search in other than plain textual fashion by default. Neither will he understand and EXPECT the results to change based on choosing the wrong Greek text.

Are first year Greek students fussing over phileo vs agaph, or are they learning about inflected forms and potentially wanting to search on specific forms? Might they want to do a study on the specific form "ego eimi" and be confused and "having to learn a new syntax" to avoid that?

Now the argument is, that these poor first year Greek students shouldn't have to learn a complicated search syntax. (Isn't learning quotation marks a learning curve too?) Ok fine, don't force people to use a syntax, make it a checkbox or something and give me the option to turn it off, and give me my complicated syntax if I want it too, so that I can cut and paste searches between different books and get the same results, GUARANTEED. Let me do everything with complicated syntax if I want to, because it's a form of expression that is easy to save and incorporate into new searches, and by giving me the option of stating unambiguously exactly what I mean, I avoid getting the wrong results if I accidently paste it into a version that does not support that feature.

Hmmm.... I'm with the Accordance folks on this one. The standard for all Boolean searching is that quotes indicate a search for an exact match to what's contained in the quotes. That's the way it is on Google if you're looking for multiple terms. It just transfers easily, universally. There's really no new syntax to be learned in my mind.

And as for learning Greek, isn't it helpful to type in a word to search for and discover all the different forms of that word - and then begin to see the ways they are formed and learn them? I think so, anyway.

And when we get into Hebrew - boy does this make things easier!!!! Just imagine if you couldn't enter a search item without getting all the pointing correct! It would be frustrating.

Anyway, my vote is for the way things are - it just works. And since "anonymous" now understands how it works, perhaps s/he can get to the business of searching the biblical text~


The more you complain about this, the more confused I become about what exactly you're frustrated about.

If you're searching a tagged text for inflected forms, with the words enclosed in quotes, and you copy and paste that search into an untagged text, you will get the exact same results, GUARANTEED. Your complaint seems to be that you are copying a lexical search from a tagged text into an untagged text and getting different results because Accordance searches an untagged text for inflected forms whether you enclose them in quotes or not.

What you seem to want is for Accordance to give you an error message when you copy and paste a lexical search into a window containing an untagged text. I'm not sure I understand why that is preferable.

Perhaps if you gave us a specific series of steps to follow, we would be better able to understand the problem you're having. But at this point, asking for the option to have a more complicated syntax so you can get more error messages makes little sense to me.

Ultimately, every program has some kind of learning curve. Whether you're a first year Greek student or a high end scholar, I think using quotation marks to specify inflected forms represents a rather shallow learning curve.

Why not just use the contextual menus available from a right mouse click or cmd mouse click? There is no guessing about what you are searching that way. Just a thought.

Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?