From: Eric D. <ede...@sy...> - 2007-02-06 21:02:47
|
Hi Mike, thank you for your comments. In response, I would say: 1) The detailed prose describing the schema is still something that needs to be done, either before or during the April workshop. 2) Your example of the meaning of m/z should probably be handled in the ontology, because the axis is defined by an ontology keyword. The ontology also needs significant work before or during the workshop. 3) Regarding your question about the type and use of float values, this should indeed be defined. I don't think this was quite so explicitly defined in either of the original specifications, but in any case, we should be as precise as possible for dataXML. Thank you for these comments. Feel free to respond and/or send more. I will assemble a list of open issues to be discussed on the list or in Lyon and resolved before release. Regards, Eric > -----Original Message----- > From: Mike Coleman [mailto:tu...@gm...] > Sent: Tuesday, February 06, 2007 7:53 AM > To: Pierre-Alain Binz; Eric Deutsch > Cc: PSI MS Dev > Subject: Re: [Psidev-ms-dev] dataXML 0.9 ready for first feedbacks >=20 > On 2/6/07, Pierre-Alain Binz <pie...@is...> wrote: > > dataXML (as being the merge between mzData and mzXML, and becoming the > PSI > > MS data standard replacing mzData) is ready for a first feedback from > you, > > implementers and PSI MS active members. We target to finalise version > 1.0 > > right after the next PSI Spring Workshop (see below) >=20 > Looking on this page, the only normative material I'm seeing are the > two XML schema. Is there other material that I'm missing? I would > have expected to see detailed normative prose describing the meaning > and proper usage of the various elements. Without this, implementers > will probably use this schema in incompatible ways, leading to future > headaches. >=20 > Here are a couple of examples of what seems to be missing. I think I > know that the X axis for spectrum data represents M/Z values, where M > is mass in amu's and Z is the integer charge. This ought to be > specified explicitly. >=20 > As a more detailed example, I think I know that spectrum data is > represented as the base64-encoded form of IEEE float values. (This > also needs to be specified.) IEEE floats can take on several special > values like "positive infinity" and error values. Is it legal to > place these values in a dataXML file? Whether or not this is so, what > is a receiver expected to do if it encounters such values? >=20 > As an example of the kind of normative prose I'm thinking of, here is > a link to a working copy of an ANSI C standard >=20 > http://www.open-std.org/jtc1/sc22/wg14/www/docs/n1124.pdf >=20 > (Obviously, the dataXML prose would be much shorter.) >=20 > Another nice thing that the ANSI C committee produces is a rationale > document, which explains the reasoning behind many of the major > decisions taken, correlated to the specification. It would be nice to > have an mzData rationale document as well, to explain the thinking > behind its design decisions (e.g., base64/binary representation of > floats). >=20 > http://www.open-std.org/jtc1/sc22/wg14/www/C99RationaleV5.10.pdf >=20 > Regards, > Mike Coleman |