Below are some comments on the details of your proposal.
But first a general comment. Although these are mostly useful features
for a text converter, the question is whether they are basic enough to
be in our model. Some, such as insert-line-numbers, might have
additional attributes to control the numbering process in some
implementations. Others, such as file-date and file-time might have a
different arrangement of attributes in some implementations, e.g.
file-date and file-time might be a single attribute with a date format
as the value.
I am not worried by features which could get more refinement with the
addition of additional attributes. That's an easy extension, and the
idea proposed by the attribute seems basic enough. But I do get
concerned about features seem too arbitrary and would likely be
performed by an entirely different attribute or set of attributes in
other implementations. The three attributes that fall into this latter
category are: file-name, file-date and file-time. Another
implementation might give very different control over headers and
footers.
This leads me to wonder if these 3 attributes would be better handled
by an extension mechanism that allows a Printer to provide additional
attributes for the special features that it provides.
Bob Herriot
> From Keith_Carter at aussmtp.austin.ibm.com Thu Feb 27 15:00:56 1997>
>>> 5.2.8.16 insert-line-numbers (boolean)
>> This job attribute specifies that the lines should be numbered in the
> document. Line
> numbers will appear throughout the document on the left side of each
> printed line.
Please add the sentence:
The number of the first line of the first page shall be 1. The number
of the first line each page, except the first page, shall be one more than
the number of the last line of the previous page. The Printer shall not
number header and footer lines.
>> 5.2.8.17 text-border (boolean)
>> This job attribute specifies that a border should be drawn around each
> logical text
> block within the text document. The border shall be drawn throughout
> the text
> document on each page. A logical text block can be the entire page or
> a column.
>> The standard values are:
>> yes Draw a border.
> no Do not draw a border.
Rather than defining "logical text block", I would prefer that the sentence
state:
This job attribute specifies that a border shall be drawn around each
column of text on each page within the document. A Printer shall
determine the size and position of each border from the size and
position of the text column, and not from the position, length or width
of the text within the text column. For example, the border for a text
column is the same whether the text fills the column or ends part way
down the column.
>> 5.2.8.19 file-name (boolean)
>> This job attribute automatically prints the file name of a document on
> each logical
> page. The file name appears throughout the document at a location on
> the page
> determined by an implementation.
>> The standard values are:
>> yes Print the file name on each logical page.
> no Do not print the file name on each logical page.
How are you defining logical page. I would prefer the following wording:
This job attribute specifies whether the Printer shall print the file
name of a document in the header or footer of each page. An
implementation shall determine the location of the file name, i.e.
header versus footer, and its position within the header or footer.
> 5.2.8.20 file-date (boolean)
Same comment as above.
> 5.2.8.21 file-time (boolean)
Same comment as above.
This attribute and the one above raise the general question that I have
above about parametization of a filter. Another implementation might
combine file-date and file-time into a single attribute with various
formats supported.
> Let's discuss at our telecon on Friday, 2/28.
If I remember correctly, our telecon is Monday 3/3 at 10 AM PST.