[sword-devel] Global search

[DM Smith]

Karl Kleinpaste wrote:
> DM Smith <dmsmith555 at yahoo.com> writes:
>   
>> I have all modules installed, too. (I don't know how you can stand all
>> the tabs!)
>>     
>
> In the Windows UI, one can right-click the tab set to bring up a picker
> to decide which modules should show in the tabs.  E.g. in the Bible
> tabs, I allow only English, Greek, Hebrew modules to show.  That said, I
> don't use Windows much, and GS doesn't have this problem -- mod.lists
> are trees opened only on demand.
>
>   
>> 1) Searching should be restricted to targeted material. For example,
>>    Don't search German Bibles for English words. An English search
>> should be targeted to English publications.
>>     
>
> How is the UI to know that what you've typed is English or German?
>   
In BD, the only search we have today is tied to first Bible in a 
parallel view. If that Bible is English, then I'd presume that the 
search is English.

I have not figured out how to integrate search for each module type (in 
terms of placement of the functionality on the GUI), but it will 
probably be targeted to a selection, whose language is known.

>   
>>    Don't search commentaries, dictionaries and other non-bibles when
>> someone is trying to find verses.
>>     
>
> How is the UI to know you're searching for verses?
>   

As above, the user will target a search against a module or a set of 
modules. Today, in BD, we have a text box tied to the parallel view of 
Bibles where a user can enter a search request. (We have a wizard that 
will help a user formulate a query that is stuffed into this text box.) 
Because of this context, we know the user is searching for verses.

> Why should commentaries be excluded from verse-oriented searches?
>   
IMHO:
If the user is trying to find verses, they are not trying to find 
commentary on verses. These are different things. The user may want to 
do both and that's fine. But as the results are different, they should 
be distinguished.

(In BD we have an option where a user can treat a commentary as a Bible. 
And then it can be put into parallel view with other Bibles. This is 
fraught with problems but was added as a quick hack to allow users to 
index and search commentaries.)

>   
>>    Don't search non-dictionaries when doing word lookup.
>>     
>
> I can think of quite a few examples of commentary content that would
> serve better than most dictionaries when doing word lookup.
>   
I meant to say "doing a dictionary word lookup".
 From an interface perspective, I was thinking of the user 
right-clicking on a word and getting a list of word based actions, one 
of which is "Find Definitions". It would also be good to have "Find 
Discussion".

>   
>> 2) Do what the user asks and expects, but offer them more.
>>     
>
>   
>> I think we should allow users the ability to tailor the applications
>> to enhance their study path. One of my grips about software in general
>> is that software is inflexible, requiring me to adapt to it.
>>     
>
> I worry about overwhelming the user in an ocean when he wants a cupful.
> I think adding configurable extensibility would be great.  But I think
> blindly adding to what is sought is a sure recipe for annoyance, and is
> a worse form of being inflexible.
>   
Yeah, that would be a problem. I think that it is important to keep a 
simple, uncluttered interface that allows the user to easily add wanted 
or remove unwanted behaviour.

Here is what I was thinking (probably inspired by Joe):
In BD, we have a sidebar, which is hidden by default, that shows a list 
of the verses from a lookup or a search. One can click on one or more 
entries in the list to focus on just their content. The user can do a 
few other operations on the list, such as deleting a selection or 
growing the context of a selection.

This list would be replaced with a tree control. The default would be 
for the tree to be collapsed, giving roughly the same view as today.

The user could expand an entry in the list to see additional stuff (some 
of the things I mentioned earlier)

The basic idea is that if the user wants more, they have to drill down 
to get it. From a visual perspective, the interface is not cluttered and 
complexity is progressively revealed on demand.

The other thought I had, was to have a robust context menu such that if 
the user were to right click within a verse or on a selection of verses, 
the would get neat things they can do with that, such as:
Find a short prioritized list of similar verses found by searching the 
lucene index for the words/Strong's numbers in these verses.
Find all entries in all works that have a cross reference that includes 
these verses.
Find all verses in this work that have a cross reference to this verse.
....
Obviously, these would have to be shortened a little bit. :)

In Him,
    DM