Re: Comparison of plain text versus databases [was: On "write function for each reference"]

[Jean Louis]

Welcome Quiliro, how are you?

* Esteban Ordóñez <quiliro@riseup.net> [2023-01-08 16:04]:
> >> In my current situation databases are "very complex"/"too complex" and
> >> plain text is "simple"... so I'm still using an approach that is based
> >> on text files, and only that approach...
> > 
> > It is opposite. Plain text is disorganized way of keeping information
> > as compared to database systems.
> 
> I have seen this addressed before.  But I am not sure yet as to the best
> for me.

It is not choice between blue and red flowers. We speak of different
classes of data storage.

> Databases are easier for the machine to search and add.

I agree to above.

> Databases take less space.

I can't say about the above, that is questionable if databases take
less space. I have no idea about it, as I have never measured. File
system may be automatically compressed, so the databases
too. Databases may create huge indexes and take more space then
necessary only to optimize for speed of searching.

> Databases are more organized.

Not automatically organized. One can disorganize databases just as
files. Though I thinkg different classes like "plain text" and
"databases", can't be compared.

Databases offer foundation to structure information in predictable
way, but files are more flexible, one could say that files also offer
that foundation, but then it depends of the user.

> Datebeses need maintenance.

For the above I can't easily agree. I heard that argument
before. Apart from system upgrade or database upgrade, I do little or
no maintenance.

Maintaining information is same as editing information, so that can't
be what you mean. In two decades I just used database, adding records,
deleting, searching, but don't know if that is meant with maintenance
and how is maintenance supposed to be a burden.

I don't feel any burden.

I do feel burden with files, sorting, searching file, and that is why
I keep index in the database mainly, so that I forget about file
names, directories, etc.

> Plain text is easier for humans to use and search.

That is questionable, and very dependend of user. Databases are
everywhere, every search engine is a database, and majority of people
know how to search database like WWW search engine. But does majority
really know how to search text in text files? I really doubt about
it. Majority does not even use editors.

Look at all the mobile applications, tasks, notes, all that more or
less structured information stored in databases. There are fields,
columns, that is what people easier adopt to.

I don't see why is plain text easier for human to use and search.

I search database every day, every hour. Let me examine it:

I have search like:

** Search people by city (free search)
** Search people by city (select among available cities)

When selecting, I just start typing, maybe "Berl" for "Berlin" and it
will show me all people with address in Berlin. There is 240452 people
in my database with 198239 addresses, with 145959 non-empty city
fields, and there are 104 people and companies from Berlin.

It was a second to find that information, another few seconds to list
people in Berlin.

Now imagine me putting 240452 in a plain text file, and all the
addresses, so how would I quickly find who is in Berlin or not?

Again I would need to have some organizing in the file, but if people
are dispersed and their information not stored in single line, I would
have serious problems getting to that data.

Considering that majority of applications today are databases, I can't
agree that: "Plain text is easier for humans to use and search".

You probably search more databases than any text, that you even forget
that you search databases.

> Plain text does not have a rigid structure.

That is very right, and that cause problems when users start demanding
the rigid structure where it is not, like in Org mode.

> Plain text takes up more space.

I can't say, I think is it other way around. In my case I use database
to find text files, so it is addition to text files, and I believe
with all indexes, database takes more space.

> Plain text needs no maintenance.

Above is correct only with read only files. Not with dynamic,
modifiable files, demanding user's attention.

Examples:

Imagine TODO statuses, just called "ACTION" or "PENDING" and similar,
like here below.

 count | actionstatuses_name 
-------+---------------------
     2 | UPDATE PENDING
     5 | PAYMENT PENDING
     7 | TERMINATED
    10 | REPEATABLE
    19 | EQUILIBRIUM
    35 | DECISION REVERSED
    94 | INEFFECTIVE
   285 | PENDING
   344 | ACTION
   506 | COMPLETED
 49388 | 

Now imagine you have in the plain text those 285 PENDING tasks, which
you need to switch to "PAYMENT PENDING", so that may be a lot of
update and maintenance if it is plain text, but just a second if it is
database.

By all means, by using database you will have less maintenance, and
more work done, and become extremely efficient. That is why they are
made for.

> Are there any other pros and cons for each?

Main argument for databases is usefulness. They are designed to be
useful.

-- 
Jean

Take action in Free Software Foundation campaigns:
https://www.fsf.org/campaigns

In support of Richard M. Stallman
https://stallmansupport.org/