OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

sarif message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [sarif] Image files: wait just a minute here!

The unit for an svg are svg units and are mapped to physical units by the renderer. Units for pdfs are 1/72". Both can be specified as just numbers if we specify the units are the natural unit for the format.

One additional wrinkle is that pdf uses a natural Cartesian coordinate system where increasing Y is upwards, whereas svg and raster formats increasing Y is downwards.

If you said that units are the natural units and did not specify which of the points is the (top, left), which is unnecessary, then specification works with rasters, svg and pdf formats.


On 05/04/2018 01:37 PM, Larry Golding (Comcast) wrote:
Also, the reason we got hung up on non-raster images was in coming up with an acceptable set of units.

-----Original Message-----
From: Larry Golding (Comcast) <larrygolding@comcast.net>
Sent: Friday, May 4, 2018 11:34 AM
To: 'Michael Fanning' <Michael.Fanning@microsoft.com>; 'James A. Kupsch' <kupsch@cs.wisc.edu>; 'sarif@lists.oasis-open.org' <sarif@lists.oasis-open.org>
Subject: RE: [sarif] Image files: wait just a minute here!

Michael, I agree with what you say, _if_ we restrict to raster images.

Jim asked why we should restrict to raster images. My reason was that we got hung up attempting to define a useful set of units for non-raster images. Jim, is there a simple way to do this?

Remember, the original motivation for this was screenshots. From my standpoint, generalizing to raster images is just gravy -- it falls out easily. Incorporating non-raster images and IMO is not worth the trouble (we could add them in a future SARIF as a non-breaking change).

(Michael, remember we've come to a meeting of the minds on the acceptability of a SARIF 2.1 that incorporates _only_ additive, non-breaking changes.)

-----Original Message-----
From: Michael Fanning <Michael.Fanning@microsoft.com>
Sent: Friday, May 4, 2018 11:29 AM
To: James A. Kupsch <kupsch@cs.wisc.edu>; Larry Golding (Comcast) <larrygolding@comcast.net>; sarif@lists.oasis-open.org
Subject: RE: [sarif] Image files: wait just a minute here!

Based on this (Larry, sorry!), we would return to Larry's original text. For raster images, rect values are >=0 integral values, where top/left has the coordinate 0,0. For vector images, the rect properties represent top, left, right, bottom values. A consumer will need to use this information to drive whatever line/polyline primitives will render the bounding rectangle.

Jim, does that sound right? Do we need language that says, you need to interpret this data, for vector formats, according to the expected coordinate system of the format? Do we need to explicitly note we support only 2D coordinate system (no 3d bounding rectangle for us)?

-----Original Message-----
From: James A. Kupsch <kupsch@cs.wisc.edu>
Sent: Friday, May 4, 2018 11:13 AM
To: Larry Golding (Comcast) <larrygolding@comcast.net>; Michael Fanning <Michael.Fanning@microsoft.com>; sarif@lists.oasis-open.org
Subject: Re: [sarif] Image files: wait just a minute here!

If images are restricted to a raster image, then it seems that all the formats I looked at use positive integer coordinate values, where the top, left value has the coordinate (0, 0).  This is not true for vector graphics formats, where negative and decimal values are allowed.

I do not see a reason to restrict images to raster formats only.


On 05/03/2018 05:21 PM, Larry Golding (Comcast) wrote:
I don’t know. *Jim*, is it possible that now that we’ve restricted
ourselves to raster-based image formats, that we can restrict the
coordinates to non-negative integers?

*From:* Michael Fanning <Michael.Fanning@microsoft.com>
*Sent:* Thursday, May 3, 2018 2:57 PM
*To:* Larry Golding (Comcast) <larrygolding@comcast.net>;
*Subject:* RE: [sarif] Image files: wait just a minute here!

Thanks Larry. I am not a raster image format expert. Is there, in
fact, a pixel-based image format that supports negative (and/or
non-integral) pixel values?

“Pixel values *MAY*be positive of negative, depending on the natural
coordinate system of the image”

*From:* Larry Golding (Comcast) <larrygolding@comcast.net
*Sent:* Thursday, May 3, 2018 2:27 PM
*To:* Michael Fanning <Michael.Fanning@microsoft.com
<mailto:Michael.Fanning@microsoft.com>>; sarif@lists.oasis-open.org
*Subject:* RE: [sarif] Image files: wait just a minute here!

FYI, here’s the relevant text in the provisional draft:

       *3.14.5 [attachment.]rectangles property*

If the attachment is a raster-based image file (for example .png,
.jpg, .gif, or .bmp), an attachment object *MAY* contain a property
namedrectangles whose value is an array of one or more unique (§3.6.2)
rectangle objects (§3.23), each of which specifies an area of interest
within the image. These rectangle objects *SHOULD* contain a message
property (§3.23.3) so a user can understand their relevance.

If the attachment is not a raster-based image file, rectangles *SHALL*
be absent.

NOTE: If a SARIF producer wished to highlight areas in a non-raster
based image file such as .svg, or a non-image file such as .pdf, it
might export the file to a raster-based image format, attach the
exported image file, and specify the rectangles on that image.


       *3.23.2 [rectangle.]top, left, bottom, and right properties*

A rectangle object *SHALL* contain properties named top, left, bottom,
and right, each of which contains a number (as defined by [JSCHEMA01
<#JSCHEMA01>]) specifying one of the coordinates of the rectangle
within the image. These properties *SHALL*be measured in pixels. Pixel
values *SHALL*increase from left to right and from top to bottom.
Pixel values *MAY*be positive of negative, depending on the natural
coordinate system of the image.

NOTE: A number in JSON schema can take a variety of forms, including
simple integers (42) and floating point numbers (3.14).


*From:* sarif@lists.oasis-open.org <mailto:sarif@lists.oasis-open.org>
<sarif@lists.oasis-open.org <mailto:sarif@lists.oasis-open.org>> *On
Behalf Of *Larry Golding (Comcast)
*Sent:* Thursday, May 3, 2018 1:28 PM
*To:* 'Michael Fanning' <Michael.Fanning@microsoft.com
<mailto:Michael.Fanning@microsoft.com>>; sarif@lists.oasis-open.org
*Subject:* RE: [sarif] Image files: wait just a minute here!

Thanks Michael. That raises a procedural issue: according to the
minutes, the TC approved #137 with certain changes. Most of them are
still good, but *this part* of this one is not:

Jim: make values float and *change pixels to units appropriate for

I will merge the changes as we’ve discussed here: supporting only
raster-based formats and requiring units of pixels (although they are
float values, and can be signed).

Then I will add an item to the agenda for the next TC meeting to
ratify this decision.


*From:* Michael Fanning <Michael.Fanning@microsoft.com
*Sent:* Wednesday, May 2, 2018 5:18 PM
*To:* Larry Golding (Comcast) <larrygolding@comcast.net
<mailto:larrygolding@comcast.net>>; sarif@lists.oasis-open.org
*Subject:* RE: [sarif] Image files: wait just a minute here!

Ok, so we might indeed have really tied ourselves into knots. Your
original text, perhaps with the addition of emphasizing that this
construct is relevant for raster image formats only, was right
(including the use of the term pixels).

In hindsight, I personally regret my contributions to this discussion.


*From:* sarif@lists.oasis-open.org <mailto:sarif@lists.oasis-open.org>
<sarif@lists.oasis-open.org <mailto:sarif@lists.oasis-open.org>> *On
Behalf Of *Larry Golding (Comcast)
*Sent:* Wednesday, May 2, 2018 4:55 PM
*To:* sarif@lists.oasis-open.org <mailto:sarif@lists.oasis-open.org>
*Subject:* [sarif] Image files: wait just a minute here!

We got tied up in knots this morning discussing how to draw rectangles
on files like PDF and SVG. But PDF is not an image format
If a SARIF producer wants to draw rectangles on a PDF, it can take a
screenshot, or use a PDF-to-JPEG converter. Likewise, SVG (although
unquestionably an image format) is a text representation. A SARIF
producer could specify a region containing the XML elements of
interest, and a viewer that knows how to render SVG could add the
necessary decorations to highlight the area. Or again, the producer
could convert to a pixel-based image file.

We can certainly continue to talk about improvements to image-like
attachments, but I propose we limit “rectangle” support to pixel-based
image files, and suggest “take a screenshot” or “convert to JPEG or PNG”
for anything else. After all, the original motivation for what became
generic image support was precisely that: screenshots.


To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail.  Follow this link to all your TCs in OASIS at:

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]