ISSUE-2: csymbol referencing attributes

DavidCarlisle comments: I agree we should add cdname and cdbase (cdbase also to
<math>) we don\'t need cdgroup

MichaelKohlhase comments: oops, that should of course be cdbase and not cdgroup.
And I think that we will need this cdbase on all cmathML elements and <math>

StanDevitt comments: I agree that we should add the cdname and cdbase
attributes. There is one crucial point though in the wording of this section. We
should not be Forcing people to use cd\'s. Recommending is okay. The reason for
this is that for completeness, I still need to be able to refer to definitions
found in (say) Euclid\'s Elements, and I do not want to place a barrier in place
that prevents references to Euclid\'s elements unless I have first written a
content dictionary for it.

Use the content dictionary if it exists. Provide some basic ones. But leave the
referencing the system still usable even at an informal level.

DavidCarlisle comments: I agree we should add cdname and cdbase (cdbase also to
<math>) we don\'t need cdgroup

MichaelKohlhase comments: oops, that should of course be cdbase and not cdgroup.
And I think that we will need this cdbase on all cmathML elements and <math>

StanDevitt comments: I agree that we should add the cdname and cdbase
attributes. There is one crucial point though in the wording of this section. We
should not be Forcing people to use cd\'s. Recommending is okay. The reason for
this is that for completeness, I still need to be able to refer to definitions
found in (say) Euclid\'s Elements, and I do not want to place a barrier in place
that prevents references to Euclid\'s elements unless I have first written a
content dictionary for it.

Use the content dictionary if it exists. Provide some basic ones. But leave the
referencing the system still usable even at an informal level.

It seems that we have agreement on this, so I propose to close this issue.

we have long fixed this issue, closing