Re: Just the issues: encoding public ids as URNs

[Norman Walsh <ndw@nwalsh.com>]

/ John Cowan <jcowan@reutershealth.com> was heard to say:

IMHO:

| Issue 1.  How shall we encode SPACE?  As %20, or by using
| one or more of the characters from Issue 9?  Note that
| leading, trailing, and consecutive spaces are impossible
| in normalized public ids.

Spaces are fairly common, and %20 is ugly, so "+"

| Issue 2.  How shall we encode APOSTROPHE?  As %27, or by using
| one or more of the characters from Issue 9?

%27

| Issue 3. How shall we encode QUESTION MARK? As %3F, or by using
| one or more of the characters from Issue 9?

%3F

| Issue 4. How shall we encode SOLIDUS? As %2F, or by using
| one or more of the characters from Issue 9?

%2F

| Issue 5. How shall we encode NUMBER SIGN? As %23, or by using
| one or more of the characters from Issue 9?

%23

| Issue 6. How shall we encode PERCENT SIGN? As %25, or by using
| one or more of the characters from Issue 9?

%25

| Issue 7.  How shall we encode the component separator in
| FPIs?  The FPI encoding is generally "//" but in one case " ".
| We could encode it in the same way as these character
| sequences, or use a special encoding.

':' in both cases.

| Issue 8.  How shall we encode the sub-component separator
| in FPIs?  The FPI encoding is "::".  We could encode it
| in the same way as these character sequences, or use a special
| encoding.

';' or ',' but I think ';' is more aesthetic for this purpose.

| Issue 9. The characters PLUS SIGN, COMMA, COLON, EQUALS SIGN,
| SEMICOLON, COMMERCIAL AT, and DOLLAR SIGN are available for use
| as part of the encoding machinery.  If we use any of them,
| how shall we encode their appearances direcly in a public id?
| (For example, if we use COLON to help deal with any of issues
| 1-8, we need an encoding for original COLONs appearing in
| the public id.)

I've suggested that we use '+', ':', and ';', so we must have an
escaping mechanism for them when they occur literally. I propose
to add '$' to this list and require the following:

 $$ to represent a literal $
 $+ to represent a literal +
 $: to represent a literal :
 $; to represent a literal ;

None of these characters occurs frequently in FPIs (as I recall) so
this shouldn't be too painful.

                                        Be seeing you,
                                          norm

-- 
Norman.Walsh@East.Sun.COM    | It is important what a man still plans
XML Standards Engineer       | at the end. It shows the measure of
Technology Development Group | injustice in his death.--Elias Canetti
Sun Microsystems, Inc.       |