Searching for Jesus in Mac Bible Software

In a recent article, I tested differences in speed, and stumbled across questions of accuracy, in the various Mac Bible study programs available. The stimulus for this test was a comment by Charles Moore of AppleLinks comparing the speed of Bible Reader Free with that of MacSword. My test was somewhat at variance with Mr. Moore's impressions, but my test was also limited to a single word search. Mr. Moore was kind enough to clarify the kinds of real-world searches upon which his initial comparison was based:

Mr. Moore then ran searches for phrases such as "Jesus wept" and "Son of Man" and listed the time it took Bible Reader Free and MacSword to perform the searches. In each case, Bible Reader Free trounced MacSword not only in the time it took to deliver initial search results, but also in the time it took to complete the entire search. In order to get a more complete picture of relative search speed among programs, let's repeat Mr. Moore's tests using each available program.

It's important to note that Mr. Moore did not use the latest version of MacSword (1.1) in his tests, so our results may be slightly different than his. It must be remembered, however, that Mr. Moore was simply trying to confirm his previous impressions about search speed using the software he already has on hand. My own mistake of using an older version—not of a program but of the KJV text—was much more egregious (see the discussion below regarding Online Bible).

It should also be noted that Moore's tests were run on a 700 MHz G3 iBook with 640 MB of RAM. My tests were run on a 1.2 GHz G4 iBook with 768 MB of RAM. Thus, where my times differ from his, the hardware difference is likely the main reason.

Mr. Moore's first test was a search for the phrase "Jesus wept," which appears only once in the New Testament (John 11:35). He writes:

Moore raises a good point that search times can be vastly improved by narrowing one's search to a range of books, such as the New Testament. While MacSword 1.01 did not have that capability, MacSword 1.1 does. Here are the times I got when doing a search of the New Testament for the phrase "Jesus wept":

MacSword's search engine offers three different kinds of searches: "multi-word," which will find every occurrence of "Jesus" and "wept" in the same verse; "exact phrase," which will only find the phrase "Jesus wept," and a "regular expression" option which I will not go into here. Search speed varies considerably depending on which kind of search you do. It appears from the results of this test that MacSword users should stick to the "multi-word" search option . . . unless of course they need to make a cup of coffee!

If we expand our search for "Jesus wept" to include the whole Bible rather than just the New Testament, we get the following times:

In my previous article, I wrote, "in terms of sheer searching speed, Bible Reader Free is by far the slowest, but in terms of the time it takes to begin working with the results of a search, Bible Reader Free is just a hair off the lead." It now appears that I need to amend that statement. The reason Bible Reader Free appeared fast in our original search for "God" is that the first occurrence of "God" appears at the beginning of the Bible. If the word or phrase being searched for does not appear until later in the search range, the time it takes for Bible Reader Free to return any search results becomes much longer. In the case of this search for "Jesus wept," Bible Reader Free showed nothing until it had slowly cycled through every book from Genesis to John. This took 13 seconds. Consequently, users of Bible Reader Free almost have to narrow the range of books being searched in order to reduce the time it takes to perform a search.

Amazingly, there was almost no difference in the time it took MacSword to search the entire Bible rather than just the New Testament! Even though the developer of MacSword touts the addition of search ranges as making possible "faster and more accurate searching," the use of a New Testament range in this case hardly affected search speed at all.

Before we move on to the second search suggested by Mr. Moore, it is important to note that Accordance, iBible, and Online Bible all handled this search with aplomb, returning results instantly, regardless of whether the search spanned the whole Bible or just the New Testament.

Mr. Moore's second test was a phrase search for Jesus' favorite term to describe Himself: "Son of man." This is an excellent test, since as Mr. Moore noted, it "appears throughout the Bible from Genesis to Revelation." Here are the times I got for each program:

Strangely, Mr. Moore wrote to me that "Bible Reader Free whipped through the entire KJV in a lively 18 seconds and returned 197 occurances of 'Son of Man.'" While I'm not sure I would describe 18 seconds as "lively," I have no idea why Mr. Moore is getting a faster time with his 700 MHz G3 than I'm getting with my 1.2 GHz G4. I suppose it's possible that there is some setting in my configuration of Bible Reader Free that is slowing down its search times, but I double-checked every setting I could find and discovered nothing which improved the speed of this search.

It is clear, however, that MacSword is, as Mr. Moore originally observed, much slower than Bible Reader Free. If one does a "multi-word" rather than an "exact phrase" search, MacSword becomes slightly faster than Bible Reader Free, but the multi-word search is useless here, since it returns every verse that has any combination of "son," "of," and "man."

It is also notable that the search for a common phrase like "son of man" slowed iBible down considerably, though its 6 second time was still much better than that of Bible Reader Free or MacSword.

This search also brought up some of the same variance in search results that I detailed in my previous article. Accordance returned 196 occurrences of the phrase "Son of man." Bible Reader Free and Online Bible each returned 197 occurrences, because they also included the one occurrence of the possessive "Son of man's." MacSword, with its double wildcard search, found 206 verses. Interestingly, MacSword appears to place the wildcards at the beginning and end of the entire phrase, rather than at the beginning and end of each word in the phrase. Thus, to duplicate MacSword's results in Accordance, I had to use "*son of man*" rather than "*son* *of* *man*". iBible also appears to have done some kind of wildcard search, though I was unable to determine exactly how it was formulated. iBible returned 220 matches, including occurrences of the phrase "son of Manasseh."

Once again, Accordance and Online Bible returned the results of this search almost instantly.

In my previous article, I was bewildered by the unusual word and verse counts I was getting with my search for "God" in the Online Bible. I eventually discovered two reasons for it. The first was that the default method of searching included the text of the KJV's footnotes (regardless of whether footnotes were displayed in the text or checked as an option in the search dialog), while the "Count" method of searching ignored the text of the footnotes. The second reason I was getting odd counts was that the search for "God" failed to find 309 of the 310 occurrences of "GOD" in all-caps. I concluded that this was a "bug" that needed to be fixed.

Jon Glass, a long-time CMUG member who knows Online Bible extremely well, wrote me this morning to explain that the exclusion of "GOD" in all caps is a feature rather than a bug:

If you will read the OLB documentation, you will discover that this is a feature, not a bug. ;-) Seriously! The KJV translators, rather than trying to transliterate the Hebrew names, chose to use English words, but in a specific way. Part of that was capitalization, so the KJV module is set to not ignore capitalization, as these capitalizations are considered different words. The capitalized "GOD" in the pagan inscription [that is, in Acts 17:23, the one occurrence of "GOD" which Online Bible did not ignore. —Ed.], however, is both NT and Greek, but also, not part of the Hebrew naming convention, so its capitalization is ignored by OLB... So, there is a method to the madness...

Jon then quoted the pertinent section of Online Bible's documentation, which explains:

Jon then lamented: "it appears that if you use Bible texts created for OLB 3 or higher, this wonderful search feature has been removed. :-( It appears that you have an older version of the KJV text in your copy of OLB."

Jon was right on all counts. While I was careful to use the latest version of the Online Bible application, it did not occur to me to make sure I was using the latest version of Online Bible's KJV text. When I updated the text this morning and ran the search again, Online Bible reported 4,473 occurrences of "God." This is consistent with our baseline count of 4,444 occurrences of "God" (whatever its case), plus the 26 occurrences of the possessive "God's" and the three occurrences of the compound word "God-ward."

While Jon laments the fact that Online Bible no longer distinguishes between "God" and "GOD," I think the developer was right to make this change. When a user does a search for the English word "God," he expects to find every occurrence of that English word, regardless of its Hebrew original. Through the use of Strong's numbers, the user can choose to distinguish between Hebrew names for God if he wants to, but it's confusing to the average user to make that distinction for him by default.

At any rate, thank you, Jon, for bringing this to my attention, and kudos to Online Bible for making this change in version 3.

Conclusion

With the help of Charles Moore and Jon Glass, we're now developing a more complete picture of the differences among the search engines of the various Mac Bible Study Programs. Bible Reader Free does appear to be faster in most cases than MacSword—even when using MacSword's most recent update. Nevertheless, iBible appears to be significantly faster than both of those programs, while Accordance and Online Bible are faster still.

In the interest of improved performance, MacSword users should choose the "multi-word" option wherever possible, while using search ranges does not appear to speed things up at all. Bible Reader Free users should, as Mr. Moore suggests, get in the habit of using search ranges to help shorten search times.

The issue of Online Bible missing words in all caps was never a general problem, but a specific choice made by the developer to help distinguish among the Hebrew names for God. This design choice has (wisely, in my opinion) been changed in recent versions: provided users have also updated to the latest version of each Bible text.

In his weblog, Rubén Gómez of Bible Software Review interacts with my first article and comes to the following conclusion:

So, what are we to make out of all this? One of the lessons is, IMO, that we should get to know (and I mean "really know") the way a program works. The second thing is that we should always double-check our results.

In the final analysis, our continued investigation into searching speed and accuracy clearly serves to underscore the wisdom of Rubén's advice.


		Searching for Jesus in Mac Bible Software Some Clarifications and Corrections by David Lang (Posted: 7/29/04)

*Getting the Whole Picture* In a recent article, I tested differences in speed, and stumbled across questions of accuracy, in the various Mac Bible study programs available. The stimulus for this test was a comment by Charles Moore of AppleLinks comparing the speed of Bible Reader Free with that of MacSword. My test was somewhat at variance with Mr. Moore's impressions, but my test was also limited to a single word search. Mr. Moore was kind enough to clarify the kinds of real-world searches upon which his initial comparison was based: Bible Reader Free does have the happy facility of turning up results sequentially as it finds them, but that is not really what I was getting at. Search parameters have a lot to do with it, I guess. In my use of Bible software, I'm most often looking for a specific phrase, rather than just a single word. Consequently, I decided to run a somewhat different comparison test between Bible Reader Free 0.98 and MacSword 1.0.1, more closely approximating a typical search I would do. Mr. Moore then ran searches for phrases such as "Jesus wept" and "Son of Man" and listed the time it took Bible Reader Free and MacSword to perform the searches. In each case, Bible Reader Free trounced MacSword not only in the time it took to deliver initial search results, but also in the time it took to complete the entire search. In order to get a more complete picture of relative search speed among programs, let's repeat Mr. Moore's tests using each available program. It's important to note that Mr. Moore did not use the latest version of MacSword (1.1) in his tests, so our results may be slightly different than his. It must be remembered, however, that Mr. Moore was simply trying to confirm his previous impressions about search speed using the software he already has on hand. My own mistake of using an older version—not of a program but of the KJV text—was much more egregious (see the discussion below regarding Online Bible). It should also be noted that Moore's tests were run on a 700 MHz G3 iBook with 640 MB of RAM. My tests were run on a 1.2 GHz G4 iBook with 768 MB of RAM. Thus, where my times differ from his, the hardware difference is likely the main reason. *"Jesus Wept"* Mr. Moore's first test was a search for the phrase "Jesus wept," which appears only once in the New Testament (John 11:35). He writes: Bible Reader Free, configured to search the New Testament only, found it in three seconds. However, MacSword does not provide for Testament-specific searches (another reason I prefer BRF), so that wasn't exactly a fair comparison, even though it is the way I would actually use the search engine. Moore raises a good point that search times can be vastly improved by narrowing one's search to a range of books, such as the New Testament. While MacSword 1.01 did not have that capability, MacSword 1.1 does. Here are the times I got when doing a search of the New Testament for the phrase "Jesus wept": Accordance, iBible, and Online Bible: instantaneous Bible Reader Free: 2 seconds to display John 11:35, 8 seconds to finish searching the New Testament MacSword: 46 seconds (phrase search) or 9 seconds (multi-word search) MacSword's search engine offers three different kinds of searches: "multi-word," which will find every occurrence of "Jesus" and "wept" in the same verse; "exact phrase," which will only find the phrase "Jesus wept," and a "regular expression" option which I will not go into here. Search speed varies considerably depending on which kind of search you do. It appears from the results of this test that MacSword users should stick to the "multi-word" search option . . . unless of course they need to make a cup of coffee! If we expand our search for "Jesus wept" to include the whole Bible rather than just the New Testament, we get the following times: Accordance, iBible, and Online Bible: instantaneous Bible Reader Free: 13 seconds to display John 11:35, 19 seconds to finish searching the whole Bible MacSword: 47 seconds (phrase search) or 10 seconds (multi-word search) In my previous article, I wrote, "in terms of sheer searching speed, Bible Reader Free is by far the slowest, but in terms of the time it takes to begin working with the results of a search, Bible Reader Free is just a hair off the lead." It now appears that I need to amend that statement. The reason Bible Reader Free appeared fast in our original search for "God" is that the first occurrence of "God" appears at the beginning of the Bible. If the word or phrase being searched for does not appear until later in the search range, the time it takes for Bible Reader Free to return any search results becomes much longer. In the case of this search for "Jesus wept," Bible Reader Free showed nothing until it had slowly cycled through every book from Genesis to John. This took 13 seconds. Consequently, users of Bible Reader Free almost have to narrow the range of books being searched in order to reduce the time it takes to perform a search. Amazingly, there was almost no difference in the time it took MacSword to search the entire Bible rather than just the New Testament! Even though the developer of MacSword touts the addition of search ranges as making possible "faster and more accurate searching," the use of a New Testament range in this case hardly affected search speed at all. Before we move on to the second search suggested by Mr. Moore, it is important to note that Accordance, iBible, and Online Bible all handled this search with aplomb, returning results instantly, regardless of whether the search spanned the whole Bible or just the New Testament. *"Son of Man"* Mr. Moore's second test was a phrase search for Jesus' favorite term to describe Himself: "Son of man." This is an excellent test, since as Mr. Moore noted, it "appears throughout the Bible from Genesis to Revelation." Here are the times I got for each program: Accordance and Online Bible: instantaneous iBible: 6 seconds Bible Reader Free: 23 seconds MacSword: 48 seconds (phrase search) or 17 seconds (multi-word search) Strangely, Mr. Moore wrote to me that "Bible Reader Free whipped through the entire KJV in a lively 18 seconds and returned 197 occurances of 'Son of Man.'" While I'm not sure I would describe 18 seconds as "lively," I have no idea why Mr. Moore is getting a faster time with his 700 MHz G3 than I'm getting with my 1.2 GHz G4. I suppose it's possible that there is some setting in my configuration of Bible Reader Free that is slowing down its search times, but I double-checked every setting I could find and discovered nothing which improved the speed of this search. It is clear, however, that MacSword is, as Mr. Moore originally observed, much slower than Bible Reader Free. If one does a "multi-word" rather than an "exact phrase" search, MacSword becomes slightly faster than Bible Reader Free, but the multi-word search is useless here, since it returns every verse that has any combination of "son," "of," and "man." It is also notable that the search for a common phrase like "son of man" slowed iBible down considerably, though its 6 second time was still much better than that of Bible Reader Free or MacSword. This search also brought up some of the same variance in search results that I detailed in my previous article. Accordance returned 196 occurrences of the phrase "Son of man." Bible Reader Free and Online Bible each returned 197 occurrences, because they also included the one occurrence of the possessive "Son of man's." MacSword, with its double wildcard search, found 206 verses. Interestingly, MacSword appears to place the wildcards at the beginning and end of the entire phrase, rather than at the beginning and end of each word in the phrase. Thus, to duplicate MacSword's results in Accordance, I had to use "son of man" rather than "son of man". iBible also appears to have done some kind of wildcard search, though I was unable to determine exactly how it was formulated. iBible returned 220 matches, including occurrences of the phrase "son of Manasseh." Once again, Accordance and Online Bible returned the results of this search almost instantly. *It's a Feature, Not a Bug* In my previous article, I was bewildered by the unusual word and verse counts I was getting with my search for "God" in the Online Bible. I eventually discovered two reasons for it. The first was that the default method of searching included the text of the KJV's footnotes (regardless of whether footnotes were displayed in the text or checked as an option in the search dialog), while the "Count" method of searching ignored the text of the footnotes. The second reason I was getting odd counts was that the search for "God" failed to find 309 of the 310 occurrences of "GOD" in all-caps. I concluded that this was a "bug" that needed to be fixed. Jon Glass, a long-time CMUG member who knows Online Bible extremely well, wrote me this morning to explain that the exclusion of "GOD" in all caps is a feature rather than a bug: If you will read the OLB documentation, you will discover that this is a feature, not a bug. ;-) Seriously! The KJV translators, rather than trying to transliterate the Hebrew names, chose to use English words, but in a specific way. Part of that was capitalization, so the KJV module is set to not ignore capitalization, as these capitalizations are considered different words. The capitalized "GOD" in the pagan inscription [that is, in Acts 17:23, the one occurrence of "GOD" which Online Bible did not ignore. —Ed.], however, is both NT and Greek, but also, not part of the Hebrew naming convention, so its capitalization is ignored by OLB... So, there is a method to the madness... Jon then quoted the pertinent section of Online Bible's documentation, which explains: When searching, the Online Bible accepts either upper or lower case letters without distinction. An exception is made in older Bible modules released prior to version 3. In those the Old Testament names of God where "GOD" (in all capitals) is distinguished from "God" (in lower case) and "LORD" (in all capitals) is distinguished from "Lord" (in lower case). To distinguish between these with version 3 or newer Bible texts, use the appropriate Strong's numbers (GOD = 03069, 03068; God = 0430, 0410, 0433, 0426; LORD = 03068, 03050, 03069; Lord = 0136, 0113, 07991, 01376, 04756, 07229). Jon then lamented: "it appears that if you use Bible texts created for OLB 3 or higher, this wonderful search feature has been removed. :-( It appears that you have an older version of the KJV text in your copy of OLB." Jon was right on all counts. While I was careful to use the latest version of the Online Bible application, it did not occur to me to make sure I was using the latest version of Online Bible's KJV text. When I updated the text this morning and ran the search again, Online Bible reported 4,473 occurrences of "God." This is consistent with our baseline count of 4,444 occurrences of "God" (whatever its case), plus the 26 occurrences of the possessive "God's" and the three occurrences of the compound word "God-ward." While Jon laments the fact that Online Bible no longer distinguishes between "God" and "GOD," I think the developer was right to make this change. When a user does a search for the English word "God," he expects to find every occurrence of that English word, regardless of its Hebrew original. Through the use of Strong's numbers, the user can choose to distinguish between Hebrew names for God if he wants to, but it's confusing to the average user to make that distinction for him by default. At any rate, thank you, Jon, for bringing this to my attention, and kudos to Online Bible for making this change in version 3. *Conclusion* With the help of Charles Moore and Jon Glass, we're now developing a more complete picture of the differences among the search engines of the various Mac Bible Study Programs. Bible Reader Free does appear to be faster in most cases than MacSword—even when using MacSword's most recent update. Nevertheless, iBible appears to be significantly faster than both of those programs, while Accordance and Online Bible are faster still. In the interest of improved performance, MacSword users should choose the "multi-word" option wherever possible, while using search ranges does not appear to speed things up at all. Bible Reader Free users should, as Mr. Moore suggests, get in the habit of using search ranges to help shorten search times. The issue of Online Bible missing words in all caps was never a general problem, but a specific choice made by the developer to help distinguish among the Hebrew names for God. This design choice has (wisely, in my opinion) been changed in recent versions: provided users have also updated to the latest version of each Bible text. In his weblog, Rubén Gómez of Bible Software Review interacts with my first article and comes to the following conclusion: So, what are we to make out of all this? One of the lessons is, IMO, that we should get to know (and I mean "really know") the way a program works. The second thing is that we should always double-check our results. In the final analysis, our continued investigation into searching speed and accuracy clearly serves to underscore the wisdom of Rubén's advice.

			David Lang is CMUG's Content Editor. David works as a developer of Accordance Bible Software, and lives in Maitland, Florida with his wife and four children.
	Other Articles: Searching for God in Mac Bible Software: A comparison of speed and accuracy. 2004 Mac Bible Software Survey Results: What you had to say about Bible Software for the Mac. Mac Bible Software Update (3): Updates galore! Accordance 6.3, MacSword 1.1, Bible Reader Free 0.98. Getting the Bible's Number: How to Use Strong's Numbers in Mac Bible Study Software. Recent Bully Pulpit Features: Apple Should Be Nothing Like Dell: Why Apple doesn't need to appeal to everyone. How Apple Really Should Be More Like Dell: Another perspective on what Dell can teach Apple about marketing. Why Apple Should Be More Like Dell: What Dell can teach Apple about marketing. Bible Software: Business or Ministry? Cutthroat Competition? Christian Charity? or Compassionate Capitalism?