Forum Posts Feed
integrating notes database with pertinent files
< Next Topic | Back to topic list | Previous Topic >
Posted by Paul Korm
May 25, 2016 at 11:45 AM
@jimspoon wrote
>So that when issues a query, the results may contain content from a notes database, and content from files, mixed together and organized in any number of different ways.
I have always liked this concept, though devilishly difficult. Results that search “content” of files depends on file type of course. Over here that would be, mainly, plain text, Word, and PDF. On the OS X side, DEVONthink’s so-called “AI” is maybe 60% of the way to @jimspoon’s vision. (It’s not really “AI”—just a proprietary dataset indexing a concordance drawn from document text content.) Though, DEVONthink seems to be on a slow downward spiral—very little advance in features for years and a partially updated UI that when released (if ever) will be nearly a decade in the making.
I think, if hints on the preview forum are accurate, that TheBrain v9 will provide better search of “notes” and attached files.
Posted by Dr Andus
May 25, 2016 at 12:59 PM
jimspoon wrote:
>Maybe the ultimate in what I am thinking of is a merger of an outliner /
>info manager with a desktop search program. So that, not only can one
>retrieve content of external files, but one can then structure,
>categorize, that content.
There is a tension here though between data and information (i.e. processed data). Even in the thread title you refer to “pertinent files,” which presumes a judgement based on some analytical process about which file is pertinent or not.
Isn’t there a danger of duplication (or multiplication) of effort, if after you have already reviewed the contents of a file, extracted some info in the form of a note, and linked to the file from the notes database, the file’s contents turn up again in search results alongside the processed notes? It could introduce unnecessary noise into the process.
So playing devil’s advocate, I’m arguing here for a tiered process, where lack of total integration is actually a good thing, and where one would use different specialised tools for both sides of the process, reserving linking for reference, in case you ever need to go back to check the original source.
But I realise there might be genuine need for total integration for some uses. Maybe Zoot can do such a thing?
Posted by Alexander Deliyannis
May 25, 2016 at 04:36 PM
jimspoon wrote:
>Very good questions. One can imagine various degrees of integration of
>“notes” and “file content”. For me, the more integrated, the better.
>So that when issues a query, the results may contain content from a
>notes database, and content from files, mixed together and organized in
>any number of different ways.
Well, a radical solution to this is for the integration to be at the note database infrastructure level, i.e. whereby the individual notes are themselves searchable files, such as plain text. This is Outwiker’s take http://jenyay.net/Outwiker/English Keepnote also took a similar approach, but it seems no longer developed. Dokuwiki also holds its info in text files, and can be configured to run locally.
In all such cases, the desktop search programme will be able to fully index the notes along with any other files, and to provide integrated search results.
Posted by Alexander Deliyannis
May 26, 2016 at 09:52 AM
One more take: Archivarius is a very powerful indexing and search utility which can read a multitude of formats http://likasoft.com/document-search/index.shtml
This includes e-mail files, as well as some PIMs databases like Mybase.
So if you use a PIM supported by Archivarius you can concurrently search and get results from pertinent files, the information manager and email.
One complementary way to improve classifications across all repositories is to include tags _within_ the content. So a tag like “#project_X” could help the search engine fetch the relevant material wherever it may be.
Posted by jimspoon
May 26, 2016 at 03:08 PM
Thanks for all the input! Dr. Andus you are right to play the devils advocate on this one.
There might be all kinds of smaller or bigger steps toward better integration of notes with external files.
One just popped into my head, and maybe it has been done somewhere. How about a links to folders, rather than just specific files ... so that when a file is placed in that folder, the info manager is aware of it, and will return “file” results in the search results. There wouldn’t be a need to find the added files and drag them into the notes database.
Another possibility - linkage of notes database tags/folders with file system tags/folders.
I’ve been keeping an eye on Outwiker ... the performance isn’t very fast, probably because it’s written in Python. Jenyay has come out with a new version, based on an updated version of python I believe. Haven’t tried it out yet.