TsertOS©® is
our
Linux distribution for x86-based systems. TsertOS©®,
see overview
is being implemented with a new User Interface (UI)
approach to folders and files, see marketing rationale,
no start menu & no
taskbar.
The Tsert::OS©®
approach
to file organization and storage, forgoes
the notion of folders when retrieving files and instead uses the
built-in search engine, see Tsert::Search©®.
Instead of looking for files
in folders, you simply query the built-in search engine
with keywords specifying the content to be found.
The results of the search is displayed, as a list of files in which
the specified keywords were found. You then only have to click on the
appropriate file.
When saving a file, the system
automatically indexes it, and performs a
content analysis, with the NLP
engine. Content
search is made
possible,
because of that analysis. See some preliminary
results
with the following files.
Other
File-System Designs
The
database approach that Microsoft
and Apple are contemplating is too complex to be implemented in a
reasonable amount of time.
It is too
complex because of the Natural Language Processing (NLP) involved in
saving the files. The approach is to split the file into
pieces of specific content or information, and store these pieces in
the
database. This makes it easy to issue queries to retrieve all pieces
which relate to a specific content, in whichever files those pieces
were found. To retrieve a specific file, we simply issue a query on the
index identifying the file, and the offsets of the pieces/chunks
of information; then sort the offsets of the chunks to reconstitute
the file.
The easy
shortcut, which would be considered a kluge, is to plug the entire file
into a blob, scan the file using a search-engine, and call
the whole thing a database filesystem. They may want
to license our NLP engine.
Back
to Top
