[seqfan] Re: Run Superseeker on existing sequence?

Marc LeBrun mlb at well.com
Sat Aug 14 01:19:09 CEST 2010

>="Donald Alan Morrison" <donmorrison at gmail.com>

[...] Personally, I find
> JSON much more human readable than XML, and just as
easy (if not easier) to
> validate.  

I agree JSON is an excellent format--I didn't wish to imply any XML
partisanship!  In fact I've written Java & Ruby services that utilized JSON
and its simplicity and directness were very congenial.  JSON's easy
machinability and low implementation cost, compared to the usual XML
libraries, were winning.

> Also, I believe some are working on
a serialized (protobuf style)
> format to supplement the plain text
format.  If you define your stream well,
> even enabling compression can
be feasible.

Neat!  I'd be interested in hearing, off-list, from whoever's hacking this.

> XML can work too, I just think
> that people get caught up
in the schema definition and transforms, which pile
> up (snowball),
rather than focusing on the lightweight problem at hand.

Yes, schema-polishing can become a distraction.  That said, there is value
in us at least informally standardizing on some things, to ease interfacing.
Shared XML schema declarations (or other codified conventions) may be useful
in applications where basic JSON doesn't quite fill the entire bill.

In any case the OEIS data is fortunately very straight-forward.  A few years
ago I had no difficulty writing a simple experimental service that scraped
the classic ASCII lookup data off the web site and served it up using a
homebrew XML schema.

It was easy, although I do recall thinking that some collaboration on the
lexicon for translating the "%" prefixes etc would eventually be in order.

Anyway, kudos for the reminder that future OEIS data will likely be embodied
in a variety of formats.

More information about the SeqFan mailing list