<{~":- !!!

Page 5 of 7 (121 items) « First ... < Previous 3 4 5 6 7 Next >
This post has 120 Replies | 9 Followers

Posts 1501
Forum MVP
Fr Devin Roza | Forum Activity | Replied: Tue, Sep 27 2016 2:15 PM

Gary Mendenhall:

The only way I'm going to be able to understand how to take advantage of those functions is if I know what they do and how they work.  In other words, "what's under the hood", or in your terms, "the mechanics".  The problem I was having was mostly in the language/terms being used and my unfamiliarity.  Was hoping to gain some insight here but see now that I was wrong and need to glean from another source for knowledge on such a sophisticated subject. 

Anyway, I have been in on the pre-pub for the advanced search training videos since it was offered and eagerly waiting for it to go live.  Keep up the good work.

Blessings,

Gary

FWIW, my main motivation to do the advanced search training videos was precisely what you express here - to try to help users get to the point where they feel comfortable and confident searching, which I am convinced necessarily includes an understanding of the essentials of what's going on under the hood, of what can be done and how it works. I hope it does the trick for you! These are difficult topics, and it's important we find the right ways to explain and make them understandable.

Posts 807
Cynthia in Florida | Forum Activity | Replied: Tue, Sep 27 2016 6:04 PM

Andrew Batishko (Faithlife):

Oh, this makes me think of a new way to explain things...

Let's create a completely brand new search language (I'm sure you'll spot some similarities). In this search language we'll try to make all terms use as similar a syntax as possible.

All searches are a term followed optionally by an operator and another term followed optionally by any number of additional operator + term pairs.

Operators are things like: AND, OR, NEAR, etc.

A term looks like: {termtype termvalues}

termtype is one of the following: word, phrase, reference, highlight, label, etc. termvalues vary and depend on the termtype.

Example terms are:{Word heaven} The termvalues here are just the text of the word you are looking for.
{Phrase heavenly father} The termvalues here are the text of the phrase you are looking for.
{Reference Bible John 3:16} The termvalues here are the datatype and the reference (whose exact format is dependent on the datatype specified).
{Highlight Yellow} The termvalues here are just the name of the highlight you are looking for.
{Label Psalms} The termvalues here are the name of the label type, followed optionally by constraints on the label attributes.

This new search language behaves exactly like the existing Logos search language. However, some of the terms defined above use a simplified syntax.

Word terms are simplified so that only the text of the word is written. {Word heaven} becomes heaven.

Phrase terms are simplified so that the text is surrounded by quotes. {Phrase heavenly father} becomes "heavenly father".

Reference terms are simplified so that the reference is surround by <>. {Reference Bible John 3:16} becomes <Bible John 3:16>. In fact, for some datatypes (such as Bible) the datatype can be removed too and the term becomes <John 3:16>.

All the other terms remain the same.

OH MY GOODNESS!!!N HOLD THE PRESSES!!!!

THIS ENTIRE POST MADE SENSE TO ME!!!!  AS IN...

-----------EVERY     

---------------------SINGLE

---------------------------------WORD!!!

EUREKA,  I THINK I'VE GOT IT! (Or at least...I'm STARTING to! Smile

Cynthia

Romans 8:28-38

Posts 807
Cynthia in Florida | Forum Activity | Replied: Tue, Sep 27 2016 6:06 PM

Eli Evans (Faithlife):

Late to the party, so just a quick note for anyone who wonders if we (FL) are a) aware that search syntax is too daggum complicated and if b) we intend to do anything about it in the future: Yes and yes.

I'm saving this thread as a design reference because I think it illustrates some key issues pretty well.

Image result for praise jesus

Cynthia

Romans 8:28-38

Posts 119
Gary Mendenhall | Forum Activity | Replied: Tue, Sep 27 2016 8:40 PM

Thank you Fr Devin Roza; its worth a lot.  I'm looking forward to the training video release.  I was a little concerned with Eli Evan's post a little earlier in this thread regarding Faithlife's being aware of the need for change in the search syntax.  I may be borrowing trouble from the future but my concern is that future changes may negate some of the knowledge gleaned from your training videos.  How frustrating it would be to finally gain knowledge and aptitude with the task of proper searching technique and then soon have it all dashed away with a completely revamped system.  Knowing the inevitability and necessity of change I find it harder and harder to adapt in my older years Huh?  Well, I'm not going to let my mind wander any further into what isn't, try to stay focused on what is, and continue to look forward to your videos. Thanks again for filling this need for clarity.Smile

Blessings to you,

Gary

Posts 13220
Forum MVP
Mark Barnes | Forum Activity | Replied: Wed, Sep 28 2016 1:52 AM

Gary Mendenhall:
How frustrating it would be to finally gain knowledge and aptitude with the task of proper searching technique and then soon have it all dashed away with a completely revamped system.

I'd be 99% certain that whatever Faithlife introduce, it won't replace the existing syntax, but complement it. Or more likely, it will be a way of generating the syntax without having to remember it. Power users would be up in arms if all our existing searches stopped working.

Posts 119
Gary Mendenhall | Forum Activity | Replied: Wed, Sep 28 2016 3:22 AM

Reassuring, Thanks Mark.

Posts 791
LogosEmployee
Eli Evans (Faithlife) | Forum Activity | Replied: Wed, Sep 28 2016 10:36 AM

Mark Barnes:

Gary Mendenhall:
How frustrating it would be to finally gain knowledge and aptitude with the task of proper searching technique and then soon have it all dashed away with a completely revamped system.

I'd be 99% certain that whatever Faithlife introduce, it won't replace the existing syntax, but complement it. Or more likely, it will be a way of generating the syntax without having to remember it. Power users would be up in arms if all our existing searches stopped working.

Quite so. If you'll indulge me in a little bit of historical contextualization, this is exactly why we have <...> and {...} today.

The last time we redesigned Search was for L4. Some syntax changes happened then, but they were relatively minor. Since then, we've built upon that foundation without changing it very much.

L5 introduced several new datasets, but they weren't searchable. (Remember how much less useful Bible Sense Lexicon was when you couldn't search the Bible for a given sense?)

Part of the mandate for L6 was to make L4 and L5 datasets searchable, as well as introducing a plethora of new datasets, plus a platform for multiplying datasets like the wind (Labels and Supplemental Datasets).

To fulfill that mandate, we decided to:

  • Extend, not replace, the existing company investment in Search technology.
  • Leverage, not sweep away, the existing user investment in Search syntax.
  • Maintain our commitment to precise searching (perhaps too much at the expense of simplicity).
  • as a result, create a multitude of heterogenous search types rather than shoehorning everything into the one monolithic search type.

The result of those guidelines was We had a choice between extending the <...> syntax so that its grammar (ie, what is considered valid within the brackets) was overly complex and potentially ambiguous, or to invent some new system for embedding multiple types of search (I agree that's a good way to explain it) within the same query. So {...} was born.

So L6 represented a quantum leap forward in both search capability and search syntax complexity.

L7 made only modest/incremental improvements on that front: Better autocomplete suggestions, more cookbook entries, and expanded help documentation. We also made some baby steps toward merging Morph and Bible search types together (ie, you can now set morph scheme at the term level instead of only at the query level).

 L7's Bible Browser is, to some extent, a way to sidestep the issue by creating a simple point-and-click UI for finding things by browsing rather than by searching (by gathering rather than by hunting).

L8, I hope, goes further in the direction of Google-ness when you want it (probably in Everything and Basic search, with undecorated terms like Judah being a catch-all cross-dataset cross-datatype matching term instead of a limited text-only search that they are today) for those who are willing to sift through more results that include more false positives (and perhaps refine their query further after the fact, or click on a "Did you mean [someething more precise]" link at the top of the results) while retaining the mathematically precise searching with term decorations <{!":- !!! for those who want the ability to specify precisely what they're looking for before the fact to get a more exact set of results. (Standard disclaimers about talking about the future apply.)

We have open tasks to add more everything to the "Everything" search. (These jokes about "everything" bagels apply. Warning: Some cussing, because that's what our culture is like, sadly. http://kottke.org/11/01/the-hilarious-everything-bagel )

Additionally, we may explore more GUI-intense ways of making complex/specialist queries. The Morph Query grid-based UI that's currently in beta/Logos Now is an overture in that direction.

Posts 1881
Donnie Hale | Forum Activity | Replied: Wed, Sep 28 2016 11:12 AM

Andrew Batishko (Faithlife):

If we were to express a reference search using search extension syntax (which doesn't actually exist, but makes it easier to compare), it might look something like: {Reference <Bible John 3:16>}

You are using Reference search to search for instances of the Bible reference John 3:16.

This example, though, brings up another question. Why does "Bible John 3:16" need to be inside <> but in the previous {Highlight Yellow} example, "Yellow" does not. That aspect of difference in search syntax has been confusing to me.

Donnie

Posts 569
Schumitinu | Forum Activity | Replied: Wed, Sep 28 2016 11:17 AM

Fr Devin Roza:

I find it helpful to think of search extensions as types of searches and datatypes as types of data. I think you've got that pretty clear. But it's also important to keep in mind that even though datatypes are "types of data", there are "types of data" that are NOT datatypes!

[...]

Every search needs three items to work: (1) type of search (2) type of data (3) data.

Oftentimes one, or even two of these three items is presupposed. This causes the confusing similarities you mention.

So, the search <Person Abraham> supplies 2 and 3, but is missing 1. It works because the search tab supplies the missing (1) type of search, which would be something like {Reference} or {Tag}.

The search {Label Psalms} supplies 1 and 3, but is missing 2. It works because the search extension supplies the (2) type of data. In cases like the Highlights or Label search extension, the type of data is considered always the same, so we don't need to supply it. It's presupposed. Note that it could have been implemented differently. We could imagine the type of data also being required - something like {Highlights Style Yellow}, or {Highlights Palette Solid Colors}.

Thank you. To be honest, I understand that better than: You search for datatype references, you search using search extensions.

Thanks everybody for trying to explain this to us.

Posts 569
Schumitinu | Forum Activity | Replied: Wed, Sep 28 2016 11:21 AM

Mark Barnes:

I like this a lot, for two reasons:

1. It helps immensely to put together the right search syntax

2. It reminds me of other things and ways, or other attributes that can be searched for.

And let me give a third reason that comes to my mind: It keeps all the different searches and their syntaxes in one place

Yes

Posts 2043
LogosEmployee
Andrew Batishko (Faithlife) | Forum Activity | Replied: Wed, Sep 28 2016 11:38 AM

Donnie Hale:

Andrew Batishko (Faithlife):

If we were to express a reference search using search extension syntax (which doesn't actually exist, but makes it easier to compare), it might look something like: {Reference <Bible John 3:16>}

You are using Reference search to search for instances of the Bible reference John 3:16.

This example, though, brings up another question. Why does "Bible John 3:16" need to be inside <> but in the previous {Highlight Yellow} example, "Yellow" does not. That aspect of difference in search syntax has been confusing to me.

Yellow does not go in <> because it's not a data type reference. It's just a string, specifically the name of a highlighting style.

This post tries to explain: https://community.logos.com/forums/p/131253/854783.aspx#854783

Posts 807
Cynthia in Florida | Forum Activity | Replied: Wed, Sep 28 2016 11:38 AM

Donnie Hale:

Andrew Batishko (Faithlife):

If we were to express a reference search using search extension syntax (which doesn't actually exist, but makes it easier to compare), it might look something like: {Reference <Bible John 3:16>}

You are using Reference search to search for instances of the Bible reference John 3:16.

This example, though, brings up another question. Why does "Bible John 3:16" need to be inside <> but in the previous {Highlight Yellow} example, "Yellow" does not. That aspect of difference in search syntax has been confusing to me.

Donnie

I have that same question!

Cynthia

Romans 8:28-38

Posts 1881
Donnie Hale | Forum Activity | Replied: Wed, Sep 28 2016 11:55 AM

Fr Devin Roza:
Every search needs three items to work: (1) type of search (2) type of data (3) data.

I think this is the second most helpful point in this thread, after Mark's "search for" vs. "search using" point. It's the synopsis of Andrew's not-so-hypothetical new search syntax.

Donnie

Posts 3115
Forum MVP
PetahChristian | Forum Activity | Replied: Wed, Sep 28 2016 12:40 PM

Donnie Hale:

Fr Devin Roza:
Every search needs three items to work: (1) type of search (2) type of data (3) data.

I think this is the second most helpful point in this thread, after Mark's "search for" vs. "search using" point. It's the synopsis of Andrew's not-so-hypothetical new search syntax.

Yes

Posts 481
Richard Villanueva | Forum Activity | Replied: Wed, Sep 28 2016 12:48 PM

Donnie Hale:
I think this is the second most helpful point in this thread, after Mark's "search for" vs. "search using" point. It's the synopsis of Andrew's not-so-hypothetical new search syntax.

This thread is a MASSIVE help in understanding Logos' search infrastructure.  Even hearing Eli's explanation of the journey in how the Search terms were improved across versions is insightful and very telling for the future (L8 and future LN releases, fingers crossed).  Kudos to Francis for starting AND for Pressing the issue to create this conversation.  It's nice seeing people put to words the questions I did not know how to properly articulate. (I did not know the things that I did not know. LOL.)

These posts have made me begin to feel empowered to create things like labels, tags, highlights, etc. in my books and resources.  I feel like I can now PURPOSEFULLY use them in order to have access to them at later dates.  Keep it coming!!

MBPro'12 / i5 / 8GB // 3.0 Scholars (Purple) / L6 & L7 Platinum, M&E Platinum, Anglican Bronze, P&C Silver / L8 Platinum, Academic Pro

Posts 1332
PL | Forum Activity | Replied: Wed, Sep 28 2016 1:26 PM

Now all we need is for MJ to extract the best bits from this thread into a nice TIP OF THE DAY.

Wonder if she would keep <{~":- !!! as the subject of the TIP... :)

Peter

Posts 1881
Donnie Hale | Forum Activity | Replied: Wed, Sep 28 2016 1:56 PM

Andrew Batishko (Faithlife):
Yellow does not go in <> because it's not a data type reference. It's just a string, specifically the name of a highlighting style.

Or... Is it a data type reference to a "Text" data type, the equivalent of {Highlight <Text Yellow>}? (I hope I'm close to right here - along the lines of Andrew's not-so-hypothetical new search syntax... ;)

Donnie

Posts 2043
LogosEmployee

Donnie Hale:

Andrew Batishko (Faithlife):
Yellow does not go in <> because it's not a data type reference. It's just a string, specifically the name of a highlighting style.

Or... Is it a data type reference to a "Text" data type, the equivalent of {Highlight <Text Yellow>}? (I hope I'm close to right here - along the lines of Andrew's not-so-hypothetical new search syntax... ;)

Not really equivalent. If you want to compare things, it would be (note that the right side below is imaginary):

<Bible John 3:16> is like {Reference WHERE DataType=Bible AND Reference=John 3:16}
{Highlight Yellow} is like {Highlight WHERE Name=Yellow}

Yellow is a Name. John 3:16 is a Reference. They are attributes of the data, and any further similarity is going to be coincidence.

Posts 26029
Forum MVP
MJ. Smith | Forum Activity | Replied: Wed, Sep 28 2016 10:34 PM

PL:
Now all we need is for MJ to extract the best bits from this thread into a nice TIP OF THE DAY.

Some was included in the TIP of the day: Best answers of the week. More will be revealed within the next few days depending upon how much time I have available ..

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Posts 13220
Forum MVP
Mark Barnes | Forum Activity | Replied: Thu, Sep 29 2016 12:23 AM

Andrew Batishko (Faithlife):
Yellow does not go in <> because it's not a data type reference.

This really matters. For anyone trying to reformulate what's Andrew/Bradley/I have said, you can't avoid distinguishing between "datatype reference" and "not a datatype reference". That really matters.

Page 5 of 7 (121 items) « First ... < Previous 3 4 5 6 7 Next > | RSS