- From: Tandy, Jeremy <jeremy.tandy@metoffice.gov.uk>
- Date: Mon, 24 Feb 2014 17:20:13 +0000
- To: Jeni Tennison <jeni@jenitennison.com>, Ivan Herman <ivan@w3.org>
- CC: W3C CSV on the Web Working Group <public-csv-wg@w3.org>
Thanks - I've added an ISSUE to the document to that effect. -----Original Message----- From: Jeni Tennison [mailto:jeni@jenitennison.com] Sent: 24 February 2014 17:16 To: Ivan Herman; Tandy, Jeremy Cc: W3C CSV on the Web Working Group Subject: Re: Organising Requirements Right: we still need to capture the semantic requirements. I don’t think we need to change either of the quotes that you (Jeremy) brought out. My remark was purely about the description of the "R-PackagingOfMultipleTables” requirement which is phrased as a requirement for packaging rather than as a requirement for grouping tables so that annotations can be assigned to the group, which is what I think it really is. Jeni ------------------------------------------------------ From: Ivan Herman ivan@w3.org Reply: Ivan Herman ivan@w3.org Date: 24 February 2014 at 11:31:34 To: Tandy, Jeremy jeremy.tandy@metoffice.gov.uk Subject: Re: Organising Requirements > > As far as I am concerned: yes. The question is whether these cases > (which are obviously real) are also present in published CSV files. If > they are (eg, if the Statistics guys indeed publish their data in CSV > 'verbatim' from those Excel sheets) then of course it is an issue for > us. But if they are, hm, smart enough to publish CSV that, for > example, does not contain the presentation artifacts, then we can > safely ignore those cases... > > Of course... there is a remark somewhere on the use case wiki page > that says that data publishers may _not_ want to publish CSV because > those very artifacts are not available, then that is of course an > issue for us, too. That may be the case of your first example below... > > Thanks! > > Ivan > > On 24 Feb 2014, at 10:35 , Tandy, Jeremy > wrote: > > > Point accepted. > > > > I think that there are two areas of the Publication of National > Statistics use case > to modify: > > > > 1) "Several tables may appear within a single sheet; for example, > refer to labour market statistics sheet 18(1). Here we see statistics > relating to headline estimates, change on quarter and change on year. > A closer inspection indicates that the column layout is identical for > each sub-table; only the meaning of the data values changes." > > > > 2) "The layout of tabular data within each sheet is optimised > for human consumption. Presentation artifacts are typically > intermingled with the data itself making it difficult for data reusers > to extract data from these tables for further processing. > Presentation artifacts, some of which are illustrated in figure Fig. 1 > Presentation artifacts in statistics worksheet, include:" > [etc.] > > > > Can you confirm that I've interpreted your concerns correctly? > > > > Thanks, Jeremy > > > > -----Original Message----- > > From: Ivan Herman [mailto:ivan@w3.org] > > Sent: 24 February 2014 08:37 > > To: Jeni Tennison > > Cc: Tandy, Jeremy; W3C CSV on the Web Working Group > > Subject: Re: Organising Requirements > > > > > > On 23 Feb 2014, at 22:03 , Jeni Tennison > wrote: > > > >> Hi Jeremy, > >> > >> Thanks for all your work pulling together the use cases and > requirements. > >> > >> Do you think it would be useful to cluster the requirements? > Looking at them, I can see: > >> > >> * Parsing, eg requirements around recognising other delimiters > >> * Annotation Types, eg R-PrimaryKey > >> * Metadata Discovery, eg R-PackagingOfMultipleTables > >> * Applications, eg R-CsvValidation > >> * Non-Functional, eg R-ZeroEditCompatibility > >> > >> Regarding the requirement R-PackagingOfMultipleTables, > I think the requirement is to annotate a group of tables, not > necessarily to package them. In other words, a design in which there > was a metadata file that pointed to a group of tables hosted elsewhere > on the web would seem to satisfy the requirement from PublicationOfNationalStatistics: > they wouldn't necessarily need to be packaged together (eg in a zip). > >> > >> Also, FWIW, I would only take syntactic requirements from > published "CSVs", not from non-text-based formats like Excel. > So, for example, I wouldn't use the ONS Excel files as demonstrating a > requirement to have multiple tables within a single CSV file. > > > > +1 > > > > We would open the flood gates if we took the original spreadsheet > (and other) programs into account... > > > > Ivan > > > >> > >> Cheers, > >> > >> Jeni > >> -- > >> Jeni Tennison > >> http://www.jenitennison.com/ > >> > > > > > > ---- > > Ivan Herman, W3C > > Digital Publishing Activity Lead > > Home: http://www.w3.org/People/Ivan/ > > mobile: +31-641044153 > > GPG: 0x343F1A3D > > FOAF: http://www.ivan-herman.net/foaf > > > > > > > > > > > > > ---- > Ivan Herman, W3C > Digital Publishing Activity Lead > Home: http://www.w3.org/People/Ivan/ > mobile: +31-641044153 > GPG: 0x343F1A3D > FOAF: http://www.ivan-herman.net/foaf > > > > > > - signature.asc, 210 bytes -- Jeni Tennison http://www.jenitennison.com/
Received on Monday, 24 February 2014 17:20:42 UTC