thanks for the example. yes, that is the example i was thinkng of. i was asking, why does adding uniqueness do any good currently. if the user has your example as follows, it does create duplicates in the output, but i don't get why that all by itself needs fixing by default.
===vvv I may try to give the example here: <<paragraph>> Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec a diam lectus. Sed sit amet ipsum mauris. <<paragraph>> Vivamus fermentum semper porta. Nunc diam velit, adipiscing ut tristique vitae, sagittis vel odio. Maecenas convallis ullamcorper ultricies. which results in: <a id="paragraph"></a> Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec a diam lectus. Sed sit amet ipsum mauris. </p> <p> <a id="paragraph"></a> Vivamus fermentum semper porta. Nunc diam velit, adipiscing ut tristique vitae, sagittis vel odio. Maecenas convallis ullamcorper ultricies. </p> ===^^^ so the duplicates in the output there are apparently a problem. i am not sure why, unless a link needs to point to each. perhaps so that the reader of the web page can point to each? but as you say, generating the document again foils that plot. [note: i started the thread about toc links. this is a subtly different question although it overlaps.] On 5/20/21, Jean Louis <bugs@gnu.support> wrote: > * Samuel Wales <samolog...@gmail.com> [2021-05-21 01:19]: >> thanks for pointing us to this variable. >> >> docstring says "This process ensures that these values are unique and >> valid...", so it sounds like you could create non-unique or invalid >> identifiers without it. >> >> does this mean, for example, if the user exports a subtree with two >> link targets with the same user label, then if this variable is >> non-nil, then the output could include more than one link target? > > I may try to give the example here: > > <<paragraph>> > Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec a diam > lectus. Sed sit amet ipsum mauris. > > <<paragraph>> > Vivamus fermentum semper porta. Nunc diam velit, adipiscing ut > tristique vitae, sagittis vel odio. Maecenas convallis ullamcorper > ultricies. > > which results in: > > <a id="paragraph"></a> > Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec a diam > lectus. Sed sit amet ipsum mauris. > </p> > > <p> > <a id="paragraph"></a> > Vivamus fermentum semper porta. Nunc diam velit, adipiscing ut > tristique vitae, sagittis vel odio. Maecenas convallis ullamcorper > ultricies. > </p> > >> what if this var were nil [the default], my brain is not working well >> now, but it seems as if the exporter could still get confused which >> target to link to, even if it is not printing duplicatedly-named >> targets. >> >> so i am curious what the purpose of the default is? > > You already discovered the purpose: "This process ensures that these > values are unique and valid..." > > Randomly generated internal hyperlinks are not part of author's > document creation and I don't believe they can be unique across all > documents as they rely on randomity, not uniqueness, but they may be > unique in one document. The sentence should say "This process ensures > that these values are unique to specific document and valid" > > Problem with it is that those random anchors/links are random, and > that makes it a bad default for user. A user may bookmark the link > > https://www.example.com/doc#org2cf8625 with some title, but with the > next document generation same link may appear as > https://www.example.com/doc#org6ac9de0 and that means that bookmark > disappeared, at least for HTML export. > > For me personally I am editing Org text (not files) on a meta level > where all objects have its unique ID and from there I can create Org > files. Then each object of a structure of meta level Org has its > unique ID. Also the document has its unique ID. Then it becomes > possible to automatically create anchors like <<1:17311:31121>> which > are truly unique across all documents, remain immutable, and are > trackable. With "trackable" I mean that it is possible to generate a > list of "referenced by" documents, documents which hyperlink to that > anchor. > > -- > Jean > > Take action in Free Software Foundation campaigns: > https://www.fsf.org/campaigns > > Sign an open letter in support of Richard M. Stallman > https://stallmansupport.org/ > https://rms-support-letter.github.io/ > > -- The Kafka Pandemic Please learn what misopathy is. https://thekafkapandemic.blogspot.com/2013/10/why-some-diseases-are-wronged.html