Re: Trying to build a database driven website



Jerry Stuckle wrote:

That's because your first premise is 100% incorrect.


I've been out of town and I don't read news on my traveling laptop.
So I'm jumping back in a bit late.
And this thread is getting a bit tired. However, you (Jerry) have
successfully dodged my main point from the beginning. All I said, at the
start, was that there was a "valid small site niche" for a files-based CMS (or page generation tool). I never said a files-based system was "better than" anything. I said it was a valid alternative to a
database, that makes particular sense in the small site context.

The one point that does make (a lot of) sense to me is the inherent
difficulty in performing search and replace operations on html
fragments, that need to be updated. In a database context these
search and replace operations (over semi-random fragments, not
labeled columns) must be framed inside query, fetch, replace and
update sequences in a database. Over files it's so easy you type
the code out at the command line. You cannot do that with query fetch
replace and update sequences, especially when the fragments you are
working over might appear in differently named columns, perhaps even
in different tables.

Html fragments (that exist in all database driven packages I've ever seen) will contain small sub-strings that periodically need to be changed. Perhaps it's a link that has been moved, where the link in question is known to exist on multiple, indeterminate pages.
Or perhaps said html fragments are known to contain references Joe Blow
is now an MD instead of an intern. That particular chore comes up
again and again and again in my work (search for a text pattern
buried in html fragments and change it to something else).

Search and replace over html fragments stored in a database
is more cumbersome than working over files. That makes room for a
niche alternative. A reference to this sort of thing (a files-based system) was mentioned in this group about a half a year ago and it
got attacked then by the usual aggressive and pretend-to-know-everything
attackers, but it also got praised by at least one of the group's most prestigious participants.

I don't think a files-based system would be appropriate for a large
commercial site. But for read-mostly informational sites of 100 pages
or less it works well indeed for me, and I find it to be so much
quicker and easier to deal with than the larger database-driven sites,
that I also worked with for years.
.



Relevant Pages

  • Re: Trying to build a database driven website
    ... I made a terrible mistake by making representations of any kind. ... A database is a ubiquitous tool. ... As such you just store data in the database, you don't format it until you write the page. ... But if so, and if those html fragments needed updating, then a query ...
    (comp.lang.php)
  • Re: Trying to build a database driven website
    ... However, you (Jerry) have ... database, that makes particular sense in the small site context. ... Or perhaps said html fragments are known to contain references Joe Blow ...
    (comp.lang.php)
  • Re: Trying to build a database driven website
    ... Good work Jerry. ... Database-driven sites don'tstore HTML fragments, ... More likely you used an application which accessed an Oracle database ...
    (comp.lang.php)
  • Re: Trying to build a database driven website
    ... Good work Jerry. ... Database-driven sites don'tstore HTML fragments, ... More likely you used an application which accessed an Oracle database and never have done any database programming. ...
    (comp.lang.php)