Cross References

[PaoloT]

4 hours ago, MikeTO said:

if you're inserting a lot of these I can see how the Affinity approach would be more work for you

It would be twice the work (for creation, maintenance, translations). If the blocks can be initially hidden, the implicit complexity of this feature may also be hidden. For example, if each block can be added by clicking a '+' button, only the basic block will be visible at first, but adding blocks would be extremely easy.

I don't know if the system adopted in ID, inherited from FrameMaker, is really needed. All considered, I think that the only types of cross-reference I've seen in my life are of these types:

- See ["page"] [page number].

- See [title] ["on page"] [page number].

- See [paragraph number] [title] ["on page"] [page number]

- See [paragraph number] [title]

- See [title]

Paolo

 

[A_B_C]

“Super-fields” would come in super-handy. I wouldn’t consider those any more overwhelming to the user than the way in which custom path names are defined in the Export Persona. So I would suggest that you simply do the following:

• Turn the options of the “Display (cross-reference) as …” menu that is available from the Edit Cross-Reference dialog into variables
• and borrow the user interface from the Path Name Definition flyout in the Export Persona to let us build super-fields in the sense suggested by Paolo T.

The user interface and its mechanics are already there. Why not use these things for something else that would be very useful? 😀

 

[A_B_C]

I hope this illustration is clear enough to express my idea. 😀

[debraspicher]

9 minutes ago, A_B_C said:

“Super-fields” would come in super-handy. I wouldn’t consider those any more overwhelming to the user than the way in which custom path names are defined in the Export Persona. So I would suggest that you simply do the following:

• Turn the options of the “Display (cross-reference) as …” menu that is available from the Edit Cross-Reference dialog into variables
• and borrow the user interface from the Path Name Definition flyout in the Export Persona to let us build super-fields in the sense suggested by Paolo T.

The user interface and its mechanics are already there. Why not use these things for something else that would be very useful? 😀

 

The Slice variables have so much potential if we can access a library of meta-data/information from the file. This can be utilized in a script when that comes along or in a template... Also user-specified variables don't work slice-to-slice atm, so it's basically non-functional.

[A_B_C]

4 hours ago, debraspicher said:

… so it's basically non-functional.

Unfortunately, I’m not exactly sure what you mean by that. My idea was not about slices or reading meta-data from files, but about utilizing a UI logic that is already available elsewhere in the application and works great, as far as it goes. If “Page Number”, “Section Name” etc. were available as variables that could be used for building expressions just as it is possible with the path name variables in path name expressions for slices, it would be possible to recreate many of the features that are available in the cross-references systems of applications like Indesign.

Think of it from a practical perspective. As Paolo said, when you have hundreds of cross-references all of which are supposed to follow a complex pattern like 

• See chapter <chapter number>, “<chapter title>”, on page <page number> of volume <volume number>, 

you will currently have to assign four (!) cross-references to create a single (!) instance of this pattern. That means, you will not end up with hundreds of cross-references, but with 4 times as much cross-references in your document.

And from a practical point of view, it should be obvious that this is a pain:

• First off, your Cross-References Panel will become overcrowded, making navigation increasingly difficult. You will easily lose track of which entries belong to the same quadruple and represent a single instance of the cross-reference pattern above.  
• And secondly, the current non-complex approach creates a potential error source of first rank. The reason being, that for each instance of the example pattern above you will have to check four pointers to locations in your document instead of a single one, where each of the pointers can be set independently from the others. By logic and economy alone, it should be obvious that the correctness of a complex cross-reference following the pattern above should not be a function of four independent data points.

So, please, give us a variable-based system from the outset. I have the impression that the target of a cross-reference is already handled internally as a complex set of data, for otherwise the user could not switch between the “Display as …” options. But if that is the case, it should be not too difficult to expose the elements of that data set to the user in the form of variables that can be used to build complex expressions. Please consider these arguments. 😀

As a side note, it would be fantastic if each of the variables could be associated with an individual character style. This way, we could, for instance, automatically format expressions that represent chapter headings (titles) in italic style, or similar. 😀

[debraspicher]

@A_B_C

Perhaps I should've expanded on the comment, but was basically agreeing the Slices panel is a great source of inspiration for expanding the potential using variable as it is currently underutilized.

Re: non-functioning: The user-defined variables portion is in the panel on bottom right of your graphic, but it's not visible as it's in the bottom. As-it-is, even if the user defines it there, it can't be used outside the scope of that slice, so it's basically not a global variable and thus no reference can be made to it.

For example, here I defined "@randomVar", but it's not able to be used in any other sliuces than this one so other than highlighting the component in a long string maybe? to aid readability, it serves no other purpose as far as we can tell.

To add another aside to the above suggestion, if we could somehow build references using individual points of data (references, meta, file data and alike) and have that made available to the user for insertion, it could open up the possibilities to create complex strings that could be reused, re-compiled on repetition. For instance, if the above example ("See chapter <chapter number>, “<chapter title>”, on page <page number> of volume <volume number>") is repeated numerous times, that string itself could also be saved as a reference and those <var>'s could already be referenced* and work/link fine...

Hopefully that's clear. If not, it's not critical.

[A_B_C]

Ah, I see. It was just a reference problem. I wasn’t sure what you were talking about. 😀

Oh, and yes, what I forgot to mention in my suggestion above: of course, we would need to have a way to create pattern presets for cross-references. Once the user is done with creating a cross-reference pattern (expression involving variables), they would need to have the option Save Preset, just as in other parts of the application. Basically, that goes without saying. Building the pattern over and over again is not much better than building the actual cross-reference itself. That would defy the idea of a pattern. 😉

In particular, it would be helpful if we could (a) store such pattern presets with a specific document only, (b) sync these patterns across parts of a book, and also (c) store the patterns also as application presets, just as we can do with color palettes.

[Intuos5]

I really like the ideas of having better custom formatting. Styling pieces of cross-references (e.g. to give a figure title an italic character style, but leave the page number in body text style) would also benefit from having something like Grep. I use this extensively in Indesign to style specific letters, words, numbers, etc. each with their own character style.

What I dislike about the way it is now is that if you accidentally hit delete in front or backspace behind a cross-reference, the entire cross-reference is gone. It would be better to show a yellow triangle that the text was modified and to only delete the entire cross-reference when all of its characters are deleted. That way, you could also opt to make final changes to cross-references in case the available formatting options are insufficient for the final print, without completely ruining all the links you set up.

What I am also missing is a preview option, so you can actually see that what you input in the cross-reference panel is what you want to get out of it (and whether or not the character style is correct).

[A_B_C]

I like the formatting-related part of your post, but I must confess I’m not entirely sure about the deletion-based part. I fear the option to manually delete parts of a cross-reference in the text would provide another error source. Imagine you could individually delete some digits of a page number that is part of a cross-reference. What should happen in such a case when you click Update cross-references? Should the truncated page number be completed then? Or should it be left alone? To be honest, when I click Update cross-references, I expect that all of my cross-references should become complete and correct references to whatever they refer to. The reference (target) may be incorrectly chosen, but the referring expression should be correct, no?

[MikeTO]

This is very minor, but I've seen this a couple of times so I thought I'd mention it. The add and edit icons can sometimes be squished. Note the horizontal scrollbar above, the panel content is about 1px wider than the panel. I haven't yet figured out which panel is allowing this to happen.

[Intuos5]

12 hours ago, A_B_C said:

I like the formatting-related part of your post, but I must confess I’m not entirely sure about the deletion-based part. I fear the option to manually delete parts of a cross-reference in the text would provide another error source. Imagine you could individually delete some digits of a page number that is part of a cross-reference. What should happen in such a case when you click Update cross-references? Should the truncated page number be completed then? Or should it be left alone? To be honest, when I click Update cross-references, I expect that all of my cross-references should become complete and correct references to whatever they refer to. The reference (target) may be incorrectly chosen, but the referring expression should be correct, no?

The best I could think of is that a partial delete should show as preflight error. And that you could hit a pencil button to edit a cross-reference manually to add/ format text while showing the portion of the text that remains linked (like you can override masterpages). In that case, you wouldn't need Grep as much and this would simplify the workflow for everyone.

[MikeTO]

Oops, I posted this in the wrong thread. Here it is again.

This bug will result in some of my cross references being wrong. If the anchor for a cross reference is at the top of a frame that is linked to a frame on a previous page, the cross reference will be to the page of the previous frame and not of the frame containing the anchor. In the attached document, a frame on page 1 is linked to a frame on page 4. There is an anchor as the first character on page 4 and when it is cross referenced the field displays as page 1. The panels show that Publisher knows it's on page 4 prior to insertion but on page 1 after insertion.

cr test.afpub

 

[Affinity Info Bot]

An issue raised in this thread ("Cross reference displays wrong page number when anchor at top of linked frame") has now been reported to the developers by the testing team (Ref: AFB-7501). Thank you very much for reporting this issue to us.

[MikeTO]

21 hours ago, MikeTO said:

This is very minor, but I've seen this a couple of times so I thought I'd mention it. The add and edit icons can sometimes be squished. Note the horizontal scrollbar above, the panel content is about 1px wider than the panel. I haven't yet figured out which panel is allowing this to happen.

This bug with squished icons is actually the default state of the panel and you can fix the icons by adjusting the column widths. The default width of the Name column is very narrow and Value is very wide when it should really be the reverse. But it's too easy to make the columns wider than the panel which leads to more squishing.

Here's the default state of the panel.

 

[Archangel]

This is a feature we have been waiting for. Excellent.

[MikeTO]

Hi, I cross referenced my entire book in the beta today for a real-world test (book not Book). It worked well, only one new crash bug. You can double click a cross ref to edit it and it works fine in the test documents I created but when I double click a cross ref in my actual book it crashes immediately. Even if I take my 2.0.4 file and try it with a couple of cross references it will crash. Here are a couple of crash reports.

Affinity Publisher 2 Beta-2023-03-16-204928.ipsAffinity Publisher 2 Beta-2023-03-16-205132.ips

Here are some suggestions which would make adding lots of cross references faster.

Cross Reference panel should show source page number: The Cross References panel shows the cross ref name and target page number. ID shows the source page number which I believe is more useful. When I'm viewing my document and there are 10 cross references to "John Smith" I need to know which page each is on, not that they're all pointing to page 123.

Default selection for Target: This field is sticky but it would be more useful if it automatically selected the document's newest anchor because if I've just created one on my own it's likely the one I want to target, not the one I targeted last time. If you wanted to get fancy it could default to the last target except if an anchor has been created since the previous cross reference was created in which case default to the new anchor, but that's likely too complex.

Show more Target rows: It would be nice to see more than 5.5 rows in the Target list. If you're xrefing a big book using anchors this will be a very long list. I like that it's a scrolling list instead of ID's inconvenient popup but a popup does show more than 5 items so it's a wash. There would be space for a taller list if the target list was on the right and the other controls were all to its left.

Filter Text: I find this field too is too narrow. It's hard to find an anchor in the short Target list so I filter by name but it's not wide enough to display two words so it shows just the second word. e.g., If I paste  "James Williams" into the field it will show "Williams" without any indication that the content has scrolled due to the field's narrow width. I typically paste in the value rather than typing it since I've just created an anchor for it and still have the text on the clipboard.

Auto select first item in Target list: I think the first item in the list should always be selected by default. When I paste "James Williams" into Filter Text and press Tab it will show me just one unselected result which I then have to select with the mouse before pressing Return or clicking OK. If the first item in the list was always selected I could just press Return to create the cross reference without using the mouse. My anchor names are all unique so pasting in a person's name will always return a single result.

Filter Style should be sticky: I also tried xrefing with paragraphs and if you have thousands of paragraphs in your document and only want to target headings you'll be using Filter Style to narrow it down but you have to keep choosing Heading 1 or 2 which gets tedious.

Max Word Count: Perhaps rename to Maximum Words?

Duplicate anchor names: Publisher doesn't require anchors to have unique names which makes xrefing a book with duplicate anchor names more challenging. It's good that the names generated when using paragraphs as targets are numbered but perhaps the insert and rename anchor features should be modified to not allow duplicate names.

Thank you! I really like the cross references feature, I just wish I could target index markers so I don't have to create all these anchors.

[Pedrober]

It's necessary to support cross-references in IDML import, together with footnotes/endnotes.

[Andreas CH]

Einzelne Endzeichen werden nicht übernommen (rot markiert).

 

[MikeTO]

This is a bug I don't really expect you to fix.

The screen recording demonstrates the cross reference field value changing each time Update is clicked, looping between page numbers 20 and 21. If the xref value is 20 the anchor will be on page 21 and if the xref value is 21 the anchor will be on page 20, so it will always be incorrect. It happens because 20 and 21 are different widths in the chosen font which causes the anchor to shift pages when the xref is updated, invalidating the updated xref.

This looping bug happens only with proportional width numerals and with an anchor just at the edge between linked frames, most likely with a pair of pages ending in 0 and 1, so it won't be too common. I tried it in Word and it has the same bug so Microsoft must have decided to live with it. I don’t know whether Adobe solved for this case. I’ve attached Publisher and Word test docs formatted with SF Pro Display so try it on a Mac. See pages 20-21 in both.

xref flip.afpubxref flip.docx

 

[Andreas CH]

Ich bin mir nicht sicher, ob das ein Fehler ist.

Wenn die Zahl «20» steht, benötigt der Absatz 3 Zeilen. Dies führt dazu, dass die letzte Zeile automatisch auf Seite 21 verschoben wird.
Mit der Zahl «21» ist der Absatz nur 2 Zeilen lang.

Dies ist nicht auf das Zählen zurückzuführen, sondern auf die «0», die weniger breit ist als die «1».

 

I am not sure if this is a mistake.

If the number is "20", the paragraph needs 3 lines. This causes the last line to be automatically moved to page 21.
With the number "21", the paragraph is only 2 lines long.

This is not due to counting, but to the "0", which is less wide than the "1".

 

 

[MikeTO]

1 minute ago, Andreas CH said:

I am not sure if this is a mistake.

If the number is "20", the paragraph needs 3 lines. This causes the last line to be automatically moved to page 21.
With the number "21", the paragraph is only 2 lines long.

This is not due to counting, but to the "0", which is less wide than the "1".

I know why the issue is occurring, I chose this font, these two page numbers, and positioned the anchor at the break point on purpose to see if it would happen. The cross reference will always be wrong so it's a bug. There's no way for a user to get the right cross reference number other than to edit the text to avoid the scenario. So if this happens in your own book just tweak the text to make it go away.

As I said I don't expect this bug to be fixed, it's an uncommon scenario.

[Andreas CH]

Ups, sorry

Die Übersetzung war mangelhaft

[MikeTO]

1 hour ago, Andreas CH said:

Ups, sorry

Die Übersetzung war mangelhaft

Don't worry, your English is probably much better than my German. 🙂 

[joe_l]

Maybe this is already covered in this thread - apart from non-deleteable cross-references?

1) When working with a BOOK, page numbers are not being inserted.

2) What is the workflow using the option "Display as: List number"? Up to now I only get 0 working in a book.

[MikeTO]

3 hours ago, joe_l said:

2) What is the workflow using the option "Display as: List number"? Up to now I only get 0 working in a book.

If you insert an anchor into:

1. Apples<anchor>

then inserting a xref to it and choosing Display As List Number will insert a 1 for numbered paragraph #1.

3 hours ago, joe_l said:

1) When working with a BOOK, page numbers are not being inserted.

 Agreed. A xref field is being inserted but no number is shown and it's not shown when exported to PDF either, even if Update Cross References Before Output is selected.

Also the Update Cross References command in the Books panel never seems to be enabled - even if the xref is moved to another page in another chapter the command is still disabled. Perhaps this is a work in progress.