Why unscan?

[caorongjin]

I am just curious about the rationale behind having an "unscan" option in 2.2+. Additionally, I would think the "scan" option can eventually be done away with if, when manually adding a citation, the scan is automatically invoked. Is there a speed concern here?

[Eyal Redler]

The idea behind the "unscan" feature is simple. It may be more convenient to work with the temp (non formatted) citation, for example, some formats will have the formatted citations as [1], [2] etc. which is less useful then {Jones, 1967} {Redler, 2002} etc.

Regarding automatically scanning, the answer is yes. It is mainly a performance issue. Remember that the bibliography software may also not be available at all times and that the bibliography may change not only when you add a citation but also when you remove one which means that having a "live" bibliography means that we have to check removed text for citations (and cut, paste etc.). This is not to say that it is impossible, we may even implement something like that in the future but for the time being we believe that we should invest our efforts in other aspects of the program...

[danzac]

what about speeding up subsequent scanning? I like scanning often to keep the paper as WYSIWYG as possible, but subsequent scanning takes just as long as the first one.

It would also be nice if the citation was still "editable" when it has been scanned, since it is still considered an object. This is most useful when changing the pages cited.

[Eyal Redler]

what about speeding up subsequent scanning?

Hopefully we'll be able to optimize this a bit in the future but you still need to scan all the document and produce the whole bibliography and this takes time. There's also a large chunk of time that is spent in the bibliography program.

It would also be nice if the citation was still "editable" when it has been scanned, since it is still considered an object.

It is editable. Maybe you mean to say that changes need to affect subsequent scans but that would be impossible. The non formatted citation is the real deal while the formatted citation is just dead text, there is no reliable way to map changes you make in the formatted citation back into the unformatted citation.

[cyberbryce]

It is editable.

I've noticed that if I edit the citation, and then add more citations, and re-scan, it rescans this citation and deletes my manual changes. That makes editing it impractical. Any way to suppress this behavior?

Best,
Bryce

[jb]

I've noticed that if I edit the citation, and then add more citations, and re-scan, it rescans this citation and deletes my manual changes. That makes editing it impractical. Any way to suppress this behavior

I've been struggling with this same issue and suspect that the answer is No--but I'd love to hear I'm wrong.
In the meantime, the best I've found is to exlcude citations from final document (in BE you can do that in preferences) and then just type the 'citation' as regular text the way I want it to appear in the final doc. Extra work, but it gives control while keeping the ability to make a bib.
Anybody have a better idea?

James

[Eyal Redler]

I've noticed that if I edit the citation, and then add more citations, and re-scan, it rescans this citation and deletes my manual changes

I'm afraid you've missed a crucial point here.

If you double click a citation you'll notice that there are two fields. One for "Citation" and one for "Formatted citation". The citation objects contains both these fields but only shows one of them at a time.

When you scan, Mellel passes whatever there is in "Citation" to the bibliography application and then gets back the "Formatted citation" and then sets the display to show the final. Any changes you make to the formatted citation will be wiped out the next time you perform a scan because Mellel uses the "Citation" field to render the final citation.

if you want to make manual changes to the citations then you need to edit the "Citation" field you can do this by either
1. Double clicking the citation and editing the field
2. Double clicking the citation, choosing "Temporary citation" in the "Display" popup, pressing return and then editing in the document.
3. Perform an "unscan" which will set all citations to display the "Temp citation" and then you can make your changes.

[cyberbryce]

I've noticed that if I edit the citation, and then add more citations, and re-scan, it rescans this citation and deletes my manual changes

If you want to make manual changes to the citations then you need to edit the "Citation" field you can do this by either
1. Double clicking the citation and editing the field
2. Double clicking the citation, choosing "Temporary citation" in the "Display" popup, pressing return and then editing in the document.
3. Perform an "unscan" which will set all citations to display the "Temp citation" and then you can make your changes.

Right, if I understand correctly, these other avenues amount to editing the citation "field", the upper box, e.g. {foobar, 1991\my custom text here\}.

And, of course, the formatted citation (the lower box), will be overwritten during the next scan. That makes sense, since how would Mellel know that the citation did not change in between scans?

Yet, it was no coincidence that I tried to edit the lower box given that it is editable: using the lower box to tweak the formatted citation seems the easiest way to ensure that your custom formatting would be as you intended. Perhaps the lower box should only be editable when you check a checkbox, "Retain Formatting On Rescan." Then Mellel would only use the field information in the upper box to re-construct the bibliography, and it would leave the in-text formatting in place.

Anyway, I have been able to achieve any special formatting by modifying the "field" code above. Perhaps others need this more than me. The confusion was created by the editable lower box more than anything.

--Bryce

[laup]

I was among those who asked for and welcomed unscan. The reasons included:

(1) I want to be sure that if I see "Jones (1998b)" in a paragraph, it was generated by the program and the proper bibliographic item inserted. That phrase might just have been typed in directly, by me or my secretary, in which case the bib item wouldn't be generated.

2) The temp citations stand out nicely to the eyes while drafting.

(3) It is for me common to revise some of the bibliographic data-base items, in which case a new scan is desirable and perhaps desirable (especially if the date has changed or another citation to the same author and year is added or deleted).

(4) I worry about automatic scanning; it sounds like one of those Microsoft-like "helps" that might not turn out to be a help, for reasons that include Eyal's points.

[danzac]

Eyal,

Thanks for the clarifications. This is exactly what I wanted and it is already there.

[eleuteruiz]

Perhaps the lower box should only be editable when you check a checkbox, "Retain Formatting On Rescan." Then Mellel would only use the field information in the upper box to re-construct the bibliography, and it would leave the in-text formatting in place.

This suggestion of Bryce does seem interesting, and would solve some problems with formatting. I need sometimes to edit formatted text in order to solve particular needs. Now I am working with 'exclude from final document' feature + manually typing my formatted text, but it is not without problems (for instance, the 'excluded' citation is transformed into a space, which will not behave like other spaces, i.e. it will not be expanded when text gets justified).