|
[
Permalink
| « Hide
]
Patrick Durusau added a comment - 06/Jul/09 01:36 PM
Proposal filed.
I have absolutely no idea what this means:
"The <draw:measure> element represents a shape that is used to measure distances in drawings." ? It is a shape that appears in drawings to enable the measurement of distance. Not sure what more could be said. How would a _shape_ be used to measure distances? If we have to points, don't we know (orcan calculate from other measurements) the distance between those points, How does this other shape gets involved? WHy would the existence of another shaoe suddenly enable us to measure a distance ? 7/13/09 -- ad hoc decided this still needs discussion, possibly need to equate a "measure shape" and "measure line"
True, the language could be cleaned up a bit but the reason I recommended rejection is that what is a "measure line" for any given shape could be quite different.
True enough, we could define and illustrate a straight edge "measure line" that would suffice for vertical and horizontal objects that present flat edges. So, how does that work if I want to include a "measure line" with an object that is drawn using a projection where the scale differs depending on where you are on the object? Or I want the "measure line" to match the shape that is is measuring, say a circle? The point being that just as there are an infinite variety of shapes there can be an infinite variety of "measure lines" to enable a user to measure them or measure between places on those objects. Should we choose to define such objects there will always be a shape for which our predefined library is insufficient. It may be that the original intent was to define a simple vertical/horizontal "measure line" and if so, I suppose we could do that but that may or may not be compatible with how people have treated this up to this point. The problem is that what you seem to understand as a <draw measure> seems to be quite difference from what I understood in the TC meeting. (And I have no idea what a measure line is from just reading the specs.) This by itself is evidence that the specs is incomplete.
You seem to imply that the <draw measure> shape could be something different from a straight line, but in 1.1 we have: The attributes that may be associated with the <draw:measure> element are: Style, Layer, Z-Index, ID, Caption ID and Transformation - see section 9.2.15 Text anchor, table background, draw end position - see section 9.2.16 Start position End position The elements that may be contained in the <draw:measure> element are: Title (short accessible name) - see section 9.2.20. Long description (in support of accessibility) - see section 9.2.20. Event listeners - see section 9.2.21. Glue points - see section 9.2.19. Text - see section 9.2.17. I don't see how this would suffice to define a shape unless we assume that it is a straight line from start position to end position. I DO NOT BELIEVE THIS ISSUE IS RESOLVED OR FIXED.
I think the preceding comment by Andreas is very important. Here is what I understand as (one notiion of) measure lines after looking around and following the ad hoc discussion. Consider some shape (a drawing figure of some kind). Consider some two points on the figure that we might wish to show a measure for (in some dimensions. In US drafting and architectural drawings these are referred to as dimension lines). Example 1: A cross-section of a mechanical assembly in which we see a cross-section (or even 3-dimensional renderiong) of a shaft. We want to indicate the diameter of the shaft, with dimensions, perhaps even tolerances. 1.1 Find two convenient points on the cross-section that are at the intersection of a diameter line with the circumference of the shaft. 1.2 Extend two parallel lines from those convenient points in some appropriate direction that reaches an open area where a legend line is to be place. 1.3 Somewhere at the point of extension of the two parallel lines, draw an intersecting line that is perpendicular to the two parallels and intersecting them both. 1.4 An inside measure line is usually show between the two parallels (that is, it projects onto the diameter through the shaft section, in this example). 1.5 An outside measure line is usually shown as line segments outside of the two parallels with nothing between, but still parallel with the diameter that locates the start of the parallel lines). This often happens when the parallel lines are close together and there is not room for arrows and legend between them on the measure line. 1.6 The measure-line segments may have arrow-heads where they intersect the parallel lines. 1.7 Someplace along the baseline that matches the measure-line, there may be dimensional information (e.g., 1.5cm, 2.0", that sort of thing). 2. A more-common example in my experience has to do with floor plans that give dimensions (measures) of doorway openings, window openings, fixtures located on the plan, etc. Wall thicknesses are often shown with the outer-form, window openings with the inner-form. 3. I suppose one can place measure lines on maps, aerial photographs, photomicrographs, and other images where annotation by dimensions and distances/lengths matter. 4. There are many variations. For example, to specify the inside dimensions of a pipe in cross-section or 3-d (isometric) view, one might do things quite differently. I SPECULATE THAT THIS IS ONE POSSIBLE WAY THAT MEASURE LINES ARE MEANT HERE. It accounts for some of the other elements that are involved. IS THERE ANY IMPLEMENTATION OF THIS FUNCTIONALITY THAT WE SHOULD TAKE INTO CONSIDERATION AS A BASIS FOR ABSTRACTING THIS FEATURE SO THAT THE SPEC. IS WELL-DEFINED? ARE THERE DRAFTING/DRAWING NOMENCLATURE DOCUMENTS THAT APPLY TO THIS ART THAT WE CAN REFERENCE? I believe that the defect report is accurate and there needs to be some sort of nomenclature definition and illustration using figures. Here is an example of measure lines and the related connectors and legends. Several variations are shown.
(It just showed up on twitter. Was published when one of our colleagues was 11 years old. I was 20 but don't recall that issue.) http://sliderulemuseum.com/Ephemera/Playboy_April1959_SlideRuleCover.jpg The description of <draw:measure> has been reworked in ODF 1.2. See
1. I think the errata treatment is unfortunate:
""Rejected, the correction of this defect would exceed the limits of OASIS errata and therefore it will be addressed in ODF 1.2. The behavior in question is implementation dependent for purposes of ISO 26300." 2. In this case, I think you can replace this with a defect resolution in section 9.2.11, adding after the first sentence a statement to the to the effect that "Determination and realization of a <draw:measure> shape is implementation dependent." 3. I think there is also a case for repairing the first sentence. I don't think the shape is used to measure distances, I think it is used to show measurements, including dimensions, on drawings. PS: The problem with the proposal shown in this JIRA issue is that the statement says what the shape is for, but it doesn't say what the shape is (and how it interacts with other shapes). To comment on Dennis statements:
1) Errata wording quoted was already adapted. 2/3) Our graphic expert Armin Le-Grand stated that the shape is being used to give dimension and measurement at the same time. See http://www.deneba.com/support/cv9g_help/Resources/DRWPREC_9107_1_15.jpg Armin stated further that a measure line is a term of art for technical drawing (directly translated from the German term Technisches Zeichnen) Those objects - in German Bemaßungsobjekte - are well known among technical drawing and I read that they are as well covered by ISO standards: DIN 1356 "Darstellung von Linien und Schraffuren in Bauzeichnungen" partly replaced by DIN ISO 3766 und DIN ISO 4066, still it is not possible to reference to these standards in an errata, as it would narrow the existing standard, existing ODF 1.0 implementations might suddenly no longer support ODF 1.0. Even more the list of used standards is long, see http://de.wikipedia.org/wiki/Technisches_Zeichnen#Spezielle_Normen Similar as before I would give no partly fix but suggest the general solution in a future version of ODF. I have no objection to saying, in the Errata, to add statements to the effect that the provision is implementation-dependent.
That or you should say that further consideration is required (in the notes) and that resolution will be provided in a future version of the specification. Either one, but not both, aye? Errata 02 - ODF 1.2 Reconciliation:
There is extensive elaboration of this provision in ODF 1.2 making the implementation-dependent clarification in Errata 02 less necessary, although it will depend on more careful analysis to determine if there is anything essential still omitted. For example, it is not clear whether the extension lines are always perpendicular to the reference line or the vector can be at a different angle. I suspect from the description that perpendicular is the only possible answer, but it just says "The first extension line is drawn on a line as defined by the vector from the start reference point to the start point of the dimension line." But the start point of the dimension line is not specified, although its distance from the reference line appears to be. So the perpendicular vector interception with that parallel dimension line probably determines the starting point of the dimension line, not vice versa. |
|||||||||||||||||||||||||||||||||||||||||
 |
|
Bug
Applied
Major
ISO/IEC JTC 1/SC 34 N 1078 : DEFECT REPORT NUMBER JP2-27
