On 09.10.2006, at 16:53, Doug Tody wrote:
> On Mon, 9 Oct 2006, Paul Harrison wrote:
>> On 09.10.2006, at 15:51, Doug Tody wrote:
>>> This approach would make sense, as not only is it more efficient
>>> for large spectral arrays of several thousand points, it is more
>>> consistent with the other serializations, which are also array-
>>> based.
>>> Data handling and transformations between formats would thus be
>>> more straightforward.
>> I would argue that the whole point of the "pure XML" serialization
>> is that it is verbose - XML is intended to be self-describing and
>> naturally hierarchical, so expressing this serialization in a form
>> that is as close to the DM as possible would seem sensible - i.e.
>> <Points> with all their associated children rather than <Flux>
>> with an array of points <Error> with an array etc. If you want to
>> be storage efficient then just use the FITS serialization....
>
> This is the usual approach with XML, and in general I agree, but in
> this case we are talking about bulk data arrays which can be thousands
> of points in length. Ignoring for the moment efficiency concerns, it
> is just a whole lot easier to deal with them as a single vector-valued
> data element. Plus as I say, this is more consistent with the other
> serializations, facilitiating conversions, and with the likely use
> of this data in applications code - which will probably also be
> array based (all current spectral applications I am familiar with
> are array based).
>
> - Doug
Still not convinced this is an argument for making the XML serialization more vector-like, its all depends on what your idea of an array is - you could still have all the abcissae,ordinate,errors etc. for a single spectral point localized in memory and then a whole spectrum
Paul Harrison
ESO Garching
www.eso.org
Received on 2006-10-09Z17:38:15