[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [xliff] Call for Dissent was Re: [xliff] csprd01 comments 11 and 41
Hi David, Ryan, I’m also confused with approved and state. The definition of approved (as of the 6 aug editor’s draft) says “Approved - Indicates whether the holding <segment> element contains a translation suitable to be used when converting the XLIFF file to original format.” One could very well want to merge a file with not final translation. So the definition of approved itself is strange. We also have a PR for <unit> that says: “A <unit> element is considered to be translated when all its <segment> children with translate attribute not set to no have the approved attribute set to yes.” What ‘translated’ means here? Is it done? Is it ready for edit?, etc. In addition the PR itself is not associated with any explicit action. What a tool is suppose to do if all segments are approved or not? I think we have to go back to the original intent of this couple of attributes, which was how to indicate that a segment is at a stage where it doesn’t need any work anymore (and what tools do with that information is up to them). I believe that state provides such information and we could remove any confusion by just removing ‘approved’. Cheers, -ys From: xliff@lists.oasis-open.org [mailto:xliff@lists.oasis-open.org] On Behalf Of Dr. David Filip Ryan, there seems to be a misunderstanding In the current solution (as described in the CFD), The approved flag is required, state and substate are otpional. state becomes necessary only if you want to use a private substate machine. Because state has 4 values and is optional, it would be too tricky to use to define a macro state of a unit, group, etc. The merit of the four values is just to provide some basic interoperability for more complicated state machines hidden in substate, that is why it is enforced with substate. approved is a simple flag that CAN [this is intentionally not normative, just a suggestion to implementers, as we do not want to bake processing logic into the spec] be used as a readiness signal If we make it recursively inherited it will be even simpler to use as the readiness signal in the sense described in the macro state definition.. Do you want to propose a specific change text/solution. I do not think that just deleting it would be satisfactory.. I thought that using <firstterm>Translated</firstterm> was a good solution, becaus it does not really prescribe an implementation behavior. Maybe it could be moved from the processing required to the front matter definition.. Please let me know during today if possible Rgds dF Cheers dF Dr. David Filip ======================= LRC | CNGL | LT-Web | CSIS University of Limerick, Ireland telephone: +353-6120-2781 cellphone: +353-86-0222-158 facsimile: +353-6120-2734 mailto: david.filip@ul.ie On Fri, Aug 9, 2013 at 5:14 AM, Ryan King <ryanki@microsoft.com> wrote: Thanks David for the explanations. I agree with state being required, but I would personally not make approved required and I would remove the processing requirement using translate=yes and review=yes as a "macro translated state". That is just too prescriptive, especially when just state=translated can serve this purpose and even more so if we make it inheritable from higher up in the tree structure. From: Dr. David Filip Thanks, Ryan let me explain below On Fri, Aug 9, 2013 at 1:19 AM, Ryan King <ryanki@microsoft.com> wrote: Hi David, All, In principle, I agree with the PR that initial is inconsistent with approved=yes and final with approved=no. However, I think that there is some overlap in meaning between reviewed and approved=yes compared to final and approved=yes. In other words, I think reviewed and approved=yes can be enough to flag that the segment is “ready to be used”. Sure there is an overlap Since approved is just a simple flag and easy to maintain, we made it required. And approved yes is intended as always sufficient readinness signal Maybe this still needs to get more explicit in the spec. At the same time we did not want to bake to much process logic into the spec, as you warned us not to.. So people who will want to benefit from all four states eventually introduce substates for even finer process granularity will probably not use approved=yes earlier than on state=final
This is a very old (long stable) part of the spec and I don't think it should be changed. This is using translated as a macro state as per one of the initial definitions. It is actually NOT about the value of the state attribute that is even not defined on unit. This is also a reason to make the flag required and vice versa, the state cannot be used to define this macro-state because it is optional.
|
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]