Re: The wisdom of the object mentors (Was: Searching OO Associations with RDBMS Persistence Models)

Robert Martin wrote:

> You are missing the point. From the point of view of the gazillions of
> users, the FS works quite nicely. I might, for the sake of argment,
> grant you that a hierarchical file system is a hideous way to represent
> data, but that does not negate the fact that it is by far the most
> popular strategy. And there is virtually no one who wants it changed.

I for one would like them changed. Every company's file system that I've worked in is a fricken mess, often 8 or more levels deep. Trees more than 4 levels deep are a yellow alert. I would much rather have a relational system, or at least one based on set theory. However, those are harder to teach to non-techies, so we are stuck with trees: simple, but not powerful.

> >
> >> There is no hue and cry for our filesystems to suddenly be
> >> RDBs.
> >
> > Umm ... Haven't you noticed the renewed drive towards
> > searching (querying)? For example Google, Google Desktop,
> > iTunes, Spotlight, Aperture? Have you even heard of WinFS?
>
> All but the last two. itunes is clearly a good candidate for a
> relational model; though I would be very upset if they integrated my
> MP3 files into a BLOB in a table somewhere!!! Google Desktop and
> Spotlight are wonderful indexing and searching technologies. But their
> output is a list of files (or in some cases emails).

We don't know that they are files. They are files when they are downloaded locally, but perhaps they are stored in a RDB on the web server side. I've programmed dynamically generated spreadsheets before. They become files on the client side, but were dynamically generated and thus never an actual file on the server (unless maybe in the buffer, but that is not visable to us).

-T- Received on Mon Jun 12 2006 - 08:12:18 CEST