Apart from editors we also need tools to validate mmCIF files for integrity, similar to what W3C has for (x)html and css.
I've mostly dealt with mmCIF reflection files so my experience with what can go wrong is limited. So far, I encountered these 'issues' that may be flagged. 1) Data items given twice with different values. This ambiguous, I suppose most parsers will use the last value given. 2) Values that should not occur for a specific data item. E.g. 19 in _refln.status 3) Proper closing of text blocks. 4) Things that can go in one loop, should go in one loop. I've seen examples where the Fmean and sigF are in one loop and I+ and I- are in another. It's not wrong, but annoying. 5) Proper space delimited values in loops. 6) Wrapping. Should this be allowed or not? I'm not a fan... 7) Data given in plain text or in new data items, even though proper data items exists. 8) Silly data such as negative amplitudes, suspiciously high values for h,k,l (such as 999), intensities between -180 and 180 There must be more things that could be checked. Cheers, Robbie Sent from my Windows Phone -----Oorspronkelijk bericht----- Van: Ethan Merritt Verzonden: 8-8-2013 2:28 Aan: CCP4BB@JISCMAIL.AC.UK Onderwerp: Re: [ccp4bb] mmCIF as working format?
