rhodes

When we compile a database of people such as an address book, it is sensible to list them by their surnames (last names), e.g., Smith, Bill and Jones, Mary. However, when we refer to them or write about their interactions it is conventional to put their first name first, e.g., "Bill Smith interviewed Mary Jones". To satisfy both needs, we need two entries for each person, e.g., "Bill Smith" and "Smith, Bill". Imagine trying to search for someone by surname here (click the pin "TSS, students" at the top).

Since PB Thought names are fixed, I need two aliases for each person or similar object in the form of two linked thoughts, one named "Bill Smith", the other "Smith, Bill". If it were possible for Instant Search to search thought labels as well as thought names, it might be possible to overcome the two-thought solution by having one alias in the name and the other in the label. However, there's been no indication from the developers that this will ever happen.

So, every time I do an Instant Search on "Bill Smith", both aliases appear, doubling the number of results. However, if it were possible to search for phrases, I could search for "Bill Smith" or just type in "bi smi" and would not get "Smith, Bill" as well.

I know you're going to suggest I use Advanced Search for this, filtering by Name only, and using the "Exact phrase:" field. But this takes several more steps to arrange, displays too much extraneous information in its results for my liking, and even then is fairly limited in what it finds. Not being an incremental search, it works very differently from Instant Search.

Its behavior can be summed up by saying that if I use "Exact phrase:", it finds too little, and if I choose "All of the words:" it finds too much. For example, if I use "Exact phrase:", I need to type "Bill Smith", not "B* Smith", "Bi?? Smith" or "ill Smith", otherwise it finds nothing. What if "Bill" happened to be a really long word? No, Advanced search is far from ideal.

I guess I could assign different types or tags to "Bill Smith" and "Smith, Bill" and filter one or the other out using Reports. But that's time-consuming, requires me to call up a Custom report each time and and limits the configurability of Reports if I'm required to keep this filter on each time I use it.

Thus I would like to see word order in Thought names carried over into Instant Search, so that typing "bi smi" finds only Bill Smith and "smi bi" finds only "Smith, Bill" (plus single matches for any other people such as Bill Smithers). Obviously, people's names is only one application where such a feature might come in handy. Any takers?

Alan Rhodes
Quote
jostber
It would be great if a more advanced regex search engine could be built into the PB. Then the user could have full control of detailed searches. An option for modular SQL searches would also expand the search functionality greatly.

Previous posts:

http://forums.thebrain.com/post?id=2823447&highlight=regex
http://forums.thebrain.com/post?id=3508576&highlight=regex



TheBrain 8.0.2.2 Slackware 14.2 KDE 4.10.3 Java 1.8 / (Windows 7)
Quote
rhodes
Regex (REGular EXpression) searching has nothing to do with the present thread since AFAIK it is not based on incremental search, used by Instant Search.

However, since you mentioned Regex, it certainly seems to be present in dtSearch's search engine module which I suspect is the one used in PB's Extended/Advanced Search (PB's About splash screen acknowledges that it uses components from this supplier). However, for some reason the developers have chosen not to implement it; maybe they can respond to this post and explain why.

Instead of Regex we have the feature known as "word expansion", which IMO gets in the way of any search precision by retrieving too many unneeded results, and does not appear capable of being turned off. However, as a lot of effort has gone into interfacing this third party module with PB in this manner, I doubt they are in any hurry to reconfigure it for something not everyone understands how to use.

Alan Rhodes
Quote
zenrain
First, I'm definitely in support of RegEx queries. However, they can be a bit of a pain to get just right, so it would be useful to be able to save the query somehow once it's been refined.

Second, I'm not clear on why you have two contacts set up for the same person. The search phrase "bi sm" found both Bill Smith and Smith, Bill for me. I only have one thought per person, in the Bill Smith format. I tried Smith, Bill for a while but 1. it's not how I think of names when I'm trying to search, and B. it hides the last name in families where they have the same last name (which I can see would be useful to some people, but drives me up the wall).
Do you need to view or sort names in order of Last, First, as well as First Last? Just trying to get the usage scenario.

Windows 7
J-1.6.0_22
--
OSX 10.6.3
Java SE 6
Quote
rhodes

I'm definitely with you there on RegEx - nice for PB to have when one needs it and it ought to be possible to save a search - but I try to structure my Brain so that 95% of the time I find everything I need using Instant Search, not Advanced/Extended search (or RegEx); reserve those for when the going gets difficult.

As for the double thought names, my problem is this: I have a thought named People with so many names that it's worth making the first generation below People an A-Z index (thoughts named A, B, ... Z). That limits it to 26 child Thoughts. Each letter has as its children my contacts whose surnames start with that letter (e.g., S contains Smith, Bill, so I can easily find all people whose names begin with S and they are already sorted alphabetically (arranged by name).

However, Bill Smith is the "main" thought for this person, and "Smith, Bill" is what I call an "index alias", i.e., an alias reserved for use in an index, as opposed to a nickname. Bill Smith and Smith, Bill are jump-linked.

I must agree that having to create two entries when one might have sufficed is a chore, but how else to do it?

- I could create a "people" type and assign it to Bill Smith, but this would exclude giving Bill any other type, e.g., according to his occupation.

- I could add the suitably un-English suffix "ppl" to each person's name and search on "bi smi ppl", but then if I use context-sensitive naming to name Bill's descendants, the suffix finds its way into the names of other thoughts too.

- I could insert the suffix "ppl" into the thought's label, but then I'd have to use Advanced Search to find it, since Instant Search can't search labels.

- I could create a tag named "people" and tag Bill with it, but then I'd have to remember to do that to every new person added. Maybe that's better. I could use Crawl Brain, starting at People, to check periodically if any grandchildren Thoughts (actual people) are untagged.

Hmm. It may be worth switching to the "people" tag idea, if only to reduce the total number of thoughts. Any other suggestions welcome. Thanks for helping me rethink this issue.

Alan Rhodes
Quote
zenrain
I see what you mean. I use a thought type "people", because to me the thought represents the person, and then have parents to group them by contextual information (for example, Author, Friend, Contact).
The only reason I could switch to this (I used to have a people parent thought as well) was the introduction of visible Thought Types in 6.x. With that I can still pull up the Type at any point and see the whole list, but in most uses I just want to see them in their context.

The other thing is I switched the sort order in Outlook and Address Book (OSX) to be in First Last order. I did this when I realized it took me twice as long to think of the persons last name and then go to that letter, than their first name, go to that letter, and narrow it down from there. The end result was I'm used to seeing things in that order now.
Then again, I'm terrible with names.

The types suggestion may or may not fit your usage scenario, but I definitely understand the "too many children" thing. Your tag idea may work better for you.


Windows 7
J-1.6.0_22
--
OSX 10.6.3
Java SE 6
Quote
rhodes
Many thanks for your suggestions. In the end I decided on the "people" tag (#4 in my earlier post) rather than a type, which might limit me in the future. Luckily it wasn't to difficult to make the changeover. I've chosen First Last too and  my Brain is a lot lighter now!
Alan Rhodes
Quote

Newsletter Signup  Newsletter        Visit TheBrain Blog   Blog       Follow us on Twitter   Twitter       Like Us on Facebook   Facebook         Watch Us on Youtube  YouTube       

TheBrain Mind Map & Mindmapping Software     Download TheBrain Mind Mapping Software