
Hi, Thanks to Bruce end others for providing input and testing it we managed to get binary that works on apple arm. Hopefully there are not more issues as we need to make a tex live 2025 snapshot this month. Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------

«Tagged PDF» still is broken: \setupbackend[ format={pdf/a-1b:2005}, profile={default_cmyk.icc,default_rgb.icc,default_gray.icc}, intent=Coated FOGRA39 (ISO 12647-2:2004), level=0] \setupstructure[state=start,method=auto] \setupbackend[export=yes] \starttext TEST \stoptext Steffen
Am 12.02.2025 um 12:05 schrieb Hans Hagen
: Hi,
Thanks to Bruce end others for providing input and testing it we managed to get binary that works on apple arm. Hopefully there are not more issues as we need to make a tex live 2025 snapshot this month.
Hans

On 2/12/2025 3:48 PM, Steffen Wolfrum wrote:
«Tagged PDF» still is broken:
\setupbackend[ format={pdf/a-1b:2005}, profile={default_cmyk.icc,default_rgb.icc,default_gray.icc}, intent=Coated FOGRA39 (ISO 12647-2:2004), level=0]
\setupstructure[state=start,method=auto] \setupbackend[export=yes]
\starttext TEST \stoptext
define broken ... you're not enabling tagging and enabling export will disable some other features so you get a different pdf anyway Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------

Am 12.02.2025 um 16:13 schrieb Hans Hagen via ntg-context
: On 2/12/2025 3:48 PM, Steffen Wolfrum wrote:
«Tagged PDF» still is broken: \setupbackend[ format={pdf/a-1b:2005}, profile={default_cmyk.icc,default_rgb.icc,default_gray.icc}, intent=Coated FOGRA39 (ISO 12647-2:2004), level=0] \setupstructure[state=start,method=auto] \setupbackend[export=yes] \starttext TEST \stoptext
define broken ... you're not enabling tagging and enabling export will disable some other features so you get a different pdf anyway
When I run this test files with e.g. ConTeXt ver: 2022.07.06 21:42 LMTX then opening the PDF in Acrobat Reader it says: Document Properties > Tagged PDF : »Yes«. When I use the current ConText its says »No«. And the publishing houses demand Tagged PDF (because of accessibility). Steffen

On 2/13/2025 3:28 PM, Steffen Wolfrum wrote:
Am 12.02.2025 um 16:13 schrieb Hans Hagen via ntg-context
: On 2/12/2025 3:48 PM, Steffen Wolfrum wrote:
«Tagged PDF» still is broken: \setupbackend[ format={pdf/a-1b:2005}, profile={default_cmyk.icc,default_rgb.icc,default_gray.icc}, intent=Coated FOGRA39 (ISO 12647-2:2004), level=0] \setupstructure[state=start,method=auto] \setupbackend[export=yes] \starttext TEST \stoptext
define broken ... you're not enabling tagging and enabling export will disable some other features so you get a different pdf anyway
When I run this test files with e.g. ConTeXt ver: 2022.07.06 21:42 LMTX then opening the PDF in Acrobat Reader it says: Document Properties > Tagged PDF : »Yes«.
When I use the current ConText its says »No«.
And the publishing houses demand Tagged PDF (because of accessibility).
whatever that means ... for printing it doesn't matter ... for viewinfb they have to provide the tools and i bet they don't give you a full acrobat prescription anyway, you don't read what we answer ... you need to enable tagging and not confuse it with exporting .. think something \setuptagging [state=start] \setupbackend [format=ua2] \starttext TEST \stoptext instead. Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------

Am 13.02.2025 um 17:37 schrieb Hans Hagen
: On 2/13/2025 3:28 PM, Steffen Wolfrum wrote:
Am 12.02.2025 um 16:13 schrieb Hans Hagen via ntg-context
: On 2/12/2025 3:48 PM, Steffen Wolfrum wrote:
«Tagged PDF» still is broken: \setupbackend[ format={pdf/a-1b:2005}, profile={default_cmyk.icc,default_rgb.icc,default_gray.icc}, intent=Coated FOGRA39 (ISO 12647-2:2004), level=0] \setupstructure[state=start,method=auto] \setupbackend[export=yes] \starttext TEST \stoptext
define broken ... you're not enabling tagging and enabling export will disable some other features so you get a different pdf anyway When I run this test files with e.g. ConTeXt ver: 2022.07.06 21:42 LMTX then opening the PDF in Acrobat Reader it says: Document Properties > Tagged PDF : »Yes«. When I use the current ConText its says »No«. And the publishing houses demand Tagged PDF (because of accessibility).
whatever that means ... for printing it doesn't matter ... for viewinfb they have to provide the tools and i bet they don't give you a full acrobat prescription
(of course not)
anyway, you don't read what we answer ... you need to enable tagging and not confuse it with exporting .. think something
\setuptagging [state=start]
\setupbackend [format=ua2]
\starttext TEST \stoptext
instead.
Hans
Does that mean that this example should provide the urgently desired alt-text: \setuptagging [state=start] \setupbackend [format=ua2] \setupexternalfigures[location=default] \starttext \startPDFalternativetext{this is a cow} \externalfigure[cow] \stopPDFalternativetext \stoptext Really?! If so, I will send it to the publishers so they can check if it fulfills their needs. Steffen

Am 14.02.2025 um 11:36 schrieb Steffen Wolfrum
: Am 13.02.2025 um 17:37 schrieb Hans Hagen
: On 2/13/2025 3:28 PM, Steffen Wolfrum wrote:
Am 12.02.2025 um 16:13 schrieb Hans Hagen via ntg-context
: On 2/12/2025 3:48 PM, Steffen Wolfrum wrote:
«Tagged PDF» still is broken: \setupbackend[ format={pdf/a-1b:2005}, profile={default_cmyk.icc,default_rgb.icc,default_gray.icc}, intent=Coated FOGRA39 (ISO 12647-2:2004), level=0] \setupstructure[state=start,method=auto] \setupbackend[export=yes] \starttext TEST \stoptext
define broken ... you're not enabling tagging and enabling export will disable some other features so you get a different pdf anyway When I run this test files with e.g. ConTeXt ver: 2022.07.06 21:42 LMTX then opening the PDF in Acrobat Reader it says: Document Properties > Tagged PDF : »Yes«. When I use the current ConText its says »No«. And the publishing houses demand Tagged PDF (because of accessibility).
whatever that means ... for printing it doesn't matter ... for viewinfb they have to provide the tools and i bet they don't give you a full acrobat prescription
(of course not)
anyway, you don't read what we answer ... you need to enable tagging and not confuse it with exporting .. think something
\setuptagging [state=start]
\setupbackend [format=ua2]
\starttext TEST \stoptext
instead.
Hans
Does that mean that this example should provide the urgently desired alt-text:
\setuptagging [state=start]
\setupbackend [format=ua2]
\setupexternalfigures[location=default]
\starttext \startPDFalternativetext{this is a cow} \externalfigure[cow] \stopPDFalternativetext \stoptext
Really?! If so, I will send it to the publishers so they can check if it fulfills their needs.
Steffen
I just received the feedback from the publishers: there is no alternative text in this example. :o( Steffen

On 2/14/25 11:57, Steffen Wolfrum wrote:
Am 14.02.2025 um 11:36 schrieb Steffen Wolfrum
: […] Does that mean that this example should provide the urgently desired alt-text: […] Really?! If so, I will send it to the publishers so they can check if it fulfills their needs.> I just received the feedback from the publishers: there is no alternative text in this example.
Hi Steffen and Hans, there is no alternative text in the PDF code because of lpdf-ini.lmt: local f_actual_text = formatters["/Span <> BDC"] - local f_alternative_text = formatters["/Span <> BDC"] + local f_alternative_text = formatters["/Span <> BDC"] local s_actual_text_e <const> = "EMC ET" The PDF key name is "Alt" (https://opensource.adobe.com/dc-acrobat-sdk-docs/pdfstandards/PDF32000_2008....), from PDF-1.3 (https://github.com/pdf-association/arlington-pdf-model/blob/master/tsv/1.3/S...) to the latest version (https://github.com/pdf-association/arlington-pdf-model/blob/master/tsv/lates...). There seems to be no "AltText" key in the PDF specification, afaIk. Besides other questions related to the accessibility directive, it would be great to fix this issue. Many thanks for your help, Pablo

On 2/14/2025 4:50 PM, Pablo Rodriguez via ntg-context wrote:
On 2/14/25 11:57, Steffen Wolfrum wrote:
Am 14.02.2025 um 11:36 schrieb Steffen Wolfrum
: […] Does that mean that this example should provide the urgently desired alt-text: […] Really?! If so, I will send it to the publishers so they can check if it fulfills their needs.> I just received the feedback from the publishers: there is no alternative text in this example. Hi Steffen and Hans,
there is no alternative text in the PDF code because of lpdf-ini.lmt:
local f_actual_text = formatters["/Span <> BDC"]
- local f_alternative_text = formatters["/Span <> BDC"] + local f_alternative_text = formatters["/Span <> BDC"]
local s_actual_text_e <const> = "EMC ET"
The PDF key name is "Alt" (https://opensource.adobe.com/dc-acrobat-sdk-docs/pdfstandards/PDF32000_2008....), from PDF-1.3 (https://github.com/pdf-association/arlington-pdf-model/blob/master/tsv/1.3/S...) to the latest version (https://github.com/pdf-association/arlington-pdf-model/blob/master/tsv/lates...).
There seems to be no "AltText" key in the PDF specification, afaIk.
Besides other questions related to the accessibility directive, it would be great to fix this issue.
As you tested viewers before, can you check which ones do something with Alt and/or ActualText at all? Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------

On 2/14/25 17:13, Hans Hagen wrote:
On 2/14/2025 4:50 PM, Pablo Rodriguez via ntg-context wrote:
[…] The PDF key name is "Alt" […] Besides other questions related to the accessibility directive, it would be great to fix this issue.
As you tested viewers before, can you check which ones do something with Alt and/or ActualText at all?
Many thanks for your fast reply, Hans. I attach the sample file (Alt value "this is a cow", ActualText value "hoh"). I don’t have access to a Windows computer now, so I cannot test Acrobat or Chrome/Edge now. With "Alt", these viewers seem to find no "cow" (using the previous sample from Steffen): PDF.js, poppler (Evince/Okular), xpdf, mupdf-gl (and SumatraPDF). With "ActualText", the compilant viewers are mupdf-gl (plus SumatraPDF), poppler (Evince/Okular), xpdf. PDF.js has only an open issue (https://github.com/mozilla/pdf.js/issues/12237). Many thanks for your help, Pablo

On 2/14/2025 6:05 PM, Pablo Rodriguez via ntg-context wrote:
On 2/14/25 17:13, Hans Hagen wrote:
On 2/14/2025 4:50 PM, Pablo Rodriguez via ntg-context wrote:
[…] The PDF key name is "Alt" […] Besides other questions related to the accessibility directive, it would be great to fix this issue.
As you tested viewers before, can you check which ones do something with Alt and/or ActualText at all?
Many thanks for your fast reply, Hans.
I attach the sample file (Alt value "this is a cow", ActualText value "hoh").
I don’t have access to a Windows computer now, so I cannot test Acrobat or Chrome/Edge now.
With "Alt", these viewers seem to find no "cow" (using the previous sample from Steffen): PDF.js, poppler (Evince/Okular), xpdf, mupdf-gl (and SumatraPDF).
With "ActualText", the compilant viewers are mupdf-gl (plus SumatraPDF), poppler (Evince/Okular), xpdf. PDF.js has only an open issue (https://github.com/mozilla/pdf.js/issues/12237).
So should one set both? Does it hurt? I suppose one is used for speach and another for copy-paste and that viewers mix and match what they like here. Could have the same values (when we set label) You can try to uncoment in lpdf-tag: -- Alt = alternate, -- Alt = pdfunicode(label ~= "" and label or "image"), irr we commented because some validator didn't like it but maybe we should not care about that too much. (Could also be under \enableexperiment control). Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------

On 2/14/25 19:18, Hans Hagen wrote:
On 2/14/2025 6:05 PM, Pablo Rodriguez via ntg-context wrote:
[…] With "Alt", these viewers seem to find no "cow" (using the previous sample from Steffen): PDF.js, poppler (Evince/Okular), xpdf, mupdf-gl (and SumatraPDF).
With "ActualText", the compilant viewers are mupdf-gl (plus SumatraPDF), poppler (Evince/Okular), xpdf. PDF.js has only an open issue (https://github.com/mozilla/pdf.js/issues/12237).
So should one set both?
Not for the same element, my fault.
Does it hurt? I suppose one is used for speach and another for copy-paste and that viewers mix and match what they like here.
Alt is for speech and to replace non-textual elements. ActualText is intended only for textual elements (text itself or images that contain text).
Could have the same values (when we set label)
You can try to uncoment in lpdf-tag:
-- Alt = alternate,
-- Alt = pdfunicode(label ~= "" and label or "image"),
irr we commented because some validator didn't like it but maybe we should not care about that too much. (Could also be under \enableexperiment control).
I will try to test this with the validators that come from the Arlington PDF model. BTW, ActualText doesn’t work in Edge (and I guess, Chrome [I don’t have access to a computer with it installed]). Many thanks for your help, Pablo

On 2/14/2025 11:36 AM, Steffen Wolfrum wrote:
Am 13.02.2025 um 17:37 schrieb Hans Hagen
: On 2/13/2025 3:28 PM, Steffen Wolfrum wrote:
Am 12.02.2025 um 16:13 schrieb Hans Hagen via ntg-context
: On 2/12/2025 3:48 PM, Steffen Wolfrum wrote:
«Tagged PDF» still is broken: \setupbackend[ format={pdf/a-1b:2005}, profile={default_cmyk.icc,default_rgb.icc,default_gray.icc}, intent=Coated FOGRA39 (ISO 12647-2:2004), level=0] \setupstructure[state=start,method=auto] \setupbackend[export=yes] \starttext TEST \stoptext
define broken ... you're not enabling tagging and enabling export will disable some other features so you get a different pdf anyway When I run this test files with e.g. ConTeXt ver: 2022.07.06 21:42 LMTX then opening the PDF in Acrobat Reader it says: Document Properties > Tagged PDF : »Yes«. When I use the current ConText its says »No«. And the publishing houses demand Tagged PDF (because of accessibility).
whatever that means ... for printing it doesn't matter ... for viewinfb they have to provide the tools and i bet they don't give you a full acrobat prescription
(of course not)
anyway, you don't read what we answer ... you need to enable tagging and not confuse it with exporting .. think something
\setuptagging [state=start]
\setupbackend [format=ua2]
\starttext TEST \stoptext
instead.
Hans
Does that mean that this example should provide the urgently desired alt-text:
\setuptagging [state=start]
\setupbackend [format=ua2]
\setupexternalfigures[location=default]
\starttext \startPDFalternativetext{this is a cow} \externalfigure[cow] \stopPDFalternativetext \stoptext
Really?! If so, I will send it to the publishers so they can check if it fulfills their needs.
I assume you can try and did so because you indicated that your viewer gave some report. Of course you need to make sure not to introduce unwanted spaces (as in your example). We can actually put an Alt on a specific image element (by setting label={...}) but for now it's disabled for now. Last time we checked some validators didn't like it and viewerd didn't do anything useful with it so ity made no sense to waste more time on it. Just putting stuff in a pdf for the sake of having something is is tricky because who knows what future validation makes of it. It's a mess (and always has been and likely will be). Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------

Am 14.02.2025 um 11:58 schrieb Hans Hagen
: On 2/14/2025 11:36 AM, Steffen Wolfrum wrote:
Am 13.02.2025 um 17:37 schrieb Hans Hagen
: On 2/13/2025 3:28 PM, Steffen Wolfrum wrote:
Am 12.02.2025 um 16:13 schrieb Hans Hagen via ntg-context
: On 2/12/2025 3:48 PM, Steffen Wolfrum wrote:
«Tagged PDF» still is broken: \setupbackend[ format={pdf/a-1b:2005}, profile={default_cmyk.icc,default_rgb.icc,default_gray.icc}, intent=Coated FOGRA39 (ISO 12647-2:2004), level=0] \setupstructure[state=start,method=auto] \setupbackend[export=yes] \starttext TEST \stoptext
define broken ... you're not enabling tagging and enabling export will disable some other features so you get a different pdf anyway When I run this test files with e.g. ConTeXt ver: 2022.07.06 21:42 LMTX then opening the PDF in Acrobat Reader it says: Document Properties > Tagged PDF : »Yes«. When I use the current ConText its says »No«. And the publishing houses demand Tagged PDF (because of accessibility).
whatever that means ... for printing it doesn't matter ... for viewinfb they have to provide the tools and i bet they don't give you a full acrobat prescription (of course not) anyway, you don't read what we answer ... you need to enable tagging and not confuse it with exporting .. think something
\setuptagging [state=start]
\setupbackend [format=ua2]
\starttext TEST \stoptext
instead.
Hans Does that mean that this example should provide the urgently desired alt-text: \setuptagging [state=start] \setupbackend [format=ua2] \setupexternalfigures[location=default] \starttext \startPDFalternativetext{this is a cow} \externalfigure[cow] \stopPDFalternativetext \stoptext Really?! If so, I will send it to the publishers so they can check if it fulfills their needs.
I assume you can try and did so because you indicated that your viewer gave some report.
Acrobat Reader can be downloaded for free. In the Document Properties it reports only whether the PDF is a «Tagged PDF» or not. The tags themselves can only be seen by Acrobat Pro. (Due to the costs probably only publishing house work with Acrobat Pro.)
Of course you need to make sure not to introduce unwanted spaces (as in your example).
I didn't want to introduce unwanted spaces. Where are spaces in my example?
We can actually put an Alt on a specific image element (by setting label={...}) but for now it's disabled for now. Last time we checked some validators didn't like it and viewerd didn't do anything useful with it so ity made no sense to waste more time on it.
Just putting stuff in a pdf for the sake of having something is is tricky because who knows what future validation makes of it. It's a mess (and always has been and likely will be).
Strange. I don't know the situation in The Netherlands, but at least in Germany all publisher are obligated by law from 28th of June 2025 on to provide books as PDF with tags and these alternative texts. How do we manage this? Steffen

On 2/14/2025 12:29 PM, Steffen Wolfrum wrote:
Am 14.02.2025 um 11:58 schrieb Hans Hagen
: On 2/14/2025 11:36 AM, Steffen Wolfrum wrote:
Am 13.02.2025 um 17:37 schrieb Hans Hagen
: On 2/13/2025 3:28 PM, Steffen Wolfrum wrote:
Am 12.02.2025 um 16:13 schrieb Hans Hagen via ntg-context
: On 2/12/2025 3:48 PM, Steffen Wolfrum wrote: > «Tagged PDF» still is broken: > \setupbackend[ > format={pdf/a-1b:2005}, > profile={default_cmyk.icc,default_rgb.icc,default_gray.icc}, > intent=Coated FOGRA39 (ISO 12647-2:2004), > level=0] > \setupstructure[state=start,method=auto] > \setupbackend[export=yes] > \starttext > TEST > \stoptext
define broken ... you're not enabling tagging and enabling export will disable some other features so you get a different pdf anyway When I run this test files with e.g. ConTeXt ver: 2022.07.06 21:42 LMTX then opening the PDF in Acrobat Reader it says: Document Properties > Tagged PDF : »Yes«. When I use the current ConText its says »No«. And the publishing houses demand Tagged PDF (because of accessibility).
whatever that means ... for printing it doesn't matter ... for viewinfb they have to provide the tools and i bet they don't give you a full acrobat prescription (of course not) anyway, you don't read what we answer ... you need to enable tagging and not confuse it with exporting .. think something
\setuptagging [state=start]
\setupbackend [format=ua2]
\starttext TEST \stoptext
instead.
Hans Does that mean that this example should provide the urgently desired alt-text: \setuptagging [state=start] \setupbackend [format=ua2] \setupexternalfigures[location=default] \starttext \startPDFalternativetext{this is a cow} \externalfigure[cow] \stopPDFalternativetext \stoptext Really?! If so, I will send it to the publishers so they can check if it fulfills their needs.
I assume you can try and did so because you indicated that your viewer gave some report.
Acrobat Reader can be downloaded for free. In the Document Properties it reports only whether the PDF is a «Tagged PDF» or not. The tags themselves can only be seen by Acrobat Pro. (Due to the costs probably only publishing house work with Acrobat Pro.)
I rarely open Acrobat Reader ... if onlty because I then need time to recover from an acute allergy for intrusive user interface changes and / or its offer to come up with summaries etc usign some ai and i don't even know of the file isn't sent somewhere.
Of course you need to make sure not to introduce unwanted spaces (as in your example).
I didn't want to introduce unwanted spaces. Where are spaces in my example?
after } and then after ]
We can actually put an Alt on a specific image element (by setting label={...}) but for now it's disabled for now. Last time we checked some validators didn't like it and viewerd didn't do anything useful with it so ity made no sense to waste more time on it.
Just putting stuff in a pdf for the sake of having something is is tricky because who knows what future validation makes of it. It's a mess (and always has been and likely will be).
Strange. I don't know the situation in The Netherlands, but at least in Germany all publisher are obligated by law from 28th of June 2025 on to provide books as PDF with tags and these alternative texts.
Do they really have to provide books as pdf? I only buy printed books so I might be out of touch with reality (i.e. getting pdf's). Also, define publisher ... I never had a publisher ask for something tagged. If they would come up with these demands I'd ask them to put their money where their mouth is (pay for development, extra work involved, tools needed).
How do we manage this?
If 'we' means 'me' (Hans) ... well, some of these rules only apply to organizations that have a turnaround over 7 digits so I'm way of from worrying. Mikael and I looked into tagging for educational materials and stopped where it still made some sense (read: university tools are happy), and will pick up in a few years when there might be some usefull tools. Also we found out that with LLM based chatbots can interpret an untagged pdf quite well so we'll see. That said: we have tagging in mkiv ... no one really cared that much (no tools that use them) ... and in mkxl we updated it a bit (still no tools) only to find out that tagged pdf (and the standard) has become fluid, seems to evolve, is kind of unstable, is a mess ... does a conforming 2 year old doc still validate? ... one cannot expect us to keep adapting for the sake of adapting, right? As it's all volunteer work, there are more instresting typesetting related things to explore. Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------

Am 14.02.2025 um 13:26 schrieb Hans Hagen
: On 2/14/2025 12:29 PM, Steffen Wolfrum wrote:
Am 14.02.2025 um 11:58 schrieb Hans Hagen
: On 2/14/2025 11:36 AM, Steffen Wolfrum wrote:
Am 13.02.2025 um 17:37 schrieb Hans Hagen
: On 2/13/2025 3:28 PM, Steffen Wolfrum wrote:
> Am 12.02.2025 um 16:13 schrieb Hans Hagen via ntg-context
: > > On 2/12/2025 3:48 PM, Steffen Wolfrum wrote: >> «Tagged PDF» still is broken: >> \setupbackend[ >> format={pdf/a-1b:2005}, >> profile={default_cmyk.icc,default_rgb.icc,default_gray.icc}, >> intent=Coated FOGRA39 (ISO 12647-2:2004), >> level=0] >> \setupstructure[state=start,method=auto] >> \setupbackend[export=yes] >> \starttext >> TEST >> \stoptext > > define broken ... you're not enabling tagging and enabling export will disable some other features so you get a different pdf anyway When I run this test files with e.g. ConTeXt ver: 2022.07.06 21:42 LMTX then opening the PDF in Acrobat Reader it says: Document Properties > Tagged PDF : »Yes«. When I use the current ConText its says »No«. And the publishing houses demand Tagged PDF (because of accessibility). whatever that means ... for printing it doesn't matter ... for viewinfb they have to provide the tools and i bet they don't give you a full acrobat prescription (of course not) anyway, you don't read what we answer ... you need to enable tagging and not confuse it with exporting .. think something
\setuptagging [state=start]
\setupbackend [format=ua2]
\starttext TEST \stoptext
instead.
Hans Does that mean that this example should provide the urgently desired alt-text: \setuptagging [state=start] \setupbackend [format=ua2] \setupexternalfigures[location=default] \starttext \startPDFalternativetext{this is a cow} \externalfigure[cow] \stopPDFalternativetext \stoptext Really?! If so, I will send it to the publishers so they can check if it fulfills their needs.
I assume you can try and did so because you indicated that your viewer gave some report. Acrobat Reader can be downloaded for free. In the Document Properties it reports only whether the PDF is a «Tagged PDF» or not. The tags themselves can only be seen by Acrobat Pro. (Due to the costs probably only publishing house work with Acrobat Pro.)
I rarely open Acrobat Reader ... if onlty because I then need time to recover from an acute allergy for intrusive user interface changes and / or its offer to come up with summaries etc usign some ai and i don't even know of the file isn't sent somewhere.
Of course you need to make sure not to introduce unwanted spaces (as in your example). I didn't want to introduce unwanted spaces. Where are spaces in my example?
after } and then after ]
We can actually put an Alt on a specific image element (by setting label={...}) but for now it's disabled for now. Last time we checked some validators didn't like it and viewerd didn't do anything useful with it so ity made no sense to waste more time on it.
Just putting stuff in a pdf for the sake of having something is is tricky because who knows what future validation makes of it. It's a mess (and always has been and likely will be). Strange. I don't know the situation in The Netherlands, but at least in Germany all publisher are obligated by law from 28th of June 2025 on to provide books as PDF with tags and these alternative texts.
Do they really have to provide books as pdf? I only buy printed books so I might be out of touch with reality (i.e. getting pdf's). Also, define publisher ...
I never had a publisher ask for something tagged. If they would come up with these demands I'd ask them to put their money where their mouth is (pay for development, extra work involved, tools needed).
How do we manage this?
If 'we' means 'me' (Hans) ... well, some of these rules only apply to organizations that have a turnaround over 7 digits so I'm way of from worrying. Mikael and I looked into tagging for educational materials and stopped where it still made some sense (read: university tools are happy), and will pick up in a few years when there might be some usefull tools. Also we found out that with LLM based chatbots can interpret an untagged pdf quite well so we'll see.
That said: we have tagging in mkiv ... no one really cared that much (no tools that use them) ... and in mkxl we updated it a bit (still no tools) only to find out that tagged pdf (and the standard) has become fluid, seems to evolve, is kind of unstable, is a mess ... does a conforming 2 year old doc still validate? ... one cannot expect us to keep adapting for the sake of adapting, right? As it's all volunteer work, there are more instresting typesetting related things to explore.
Hi Hans, in my case the publisher that needs this feature right now is Heidelberg University Publishing: https://heiup.uni-heidelberg.de/ Definitely no turnaround over 7 digits. A small but committed publisher, that is engaged in making PDF documents also accessible for blind people. Yours, Steffen

On 2/14/2025 8:03 PM, Steffen Wolfrum wrote:
in my case the publisher that needs this feature right now is Heidelberg University Publishing: https://heiup.uni-heidelberg.de/
Definitely no turnaround over 7 digits. A small but committed publisher, that is engaged in making PDF documents also accessible for blind people.
- adding some Alt whatever to a pdf has been possible for ages but only when it's required by law they do it - adding some simple label is not helping the blind, while making a dedicated document (or providing the sources) with more extensive description of what this image is showing makes more sense - add audio resources to documents clearly explaining matters that require something that a blind person needs seems better to me (also something that can be done for ages) so ... if publishers really cared about that they'd embraced tex as tool for *multiple* output (not only blind, maybe also huge fonts etc for bad eyesight or alternative coloring for color blind or ...) not even now, with accessibility on the agenda, those demands show up ... so forgive if i don't take publishers demands too serious on this anyway, we'll see what direction pablos tests show us, and i'll adapt to that Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------

Am 14.02.2025 um 20:43 schrieb Hans Hagen
: On 2/14/2025 8:03 PM, Steffen Wolfrum wrote:
in my case the publisher that needs this feature right now is Heidelberg University Publishing: https://heiup.uni-heidelberg.de/ Definitely no turnaround over 7 digits. A small but committed publisher, that is engaged in making PDF documents also accessible for blind people.
- adding some Alt whatever to a pdf has been possible for ages but only when it's required by law they do it
- adding some simple label is not helping the blind, while making a dedicated document (or providing the sources) with more extensive description of what this image is showing makes more sense
- add audio resources to documents clearly explaining matters that require something that a blind person needs seems better to me (also something that can be done for ages)
so ... if publishers really cared about that they'd embraced tex as tool for *multiple* output (not only blind, maybe also huge fonts etc for bad eyesight or alternative coloring for color blind or ...)
not even now, with accessibility on the agenda, those demands show up ... so forgive if i don't take publishers demands too serious on this
anyway, we'll see what direction pablos tests show us, and i'll adapt to that
Hans
I agree completely. Steffen

On 2/14/25 20:03, Steffen Wolfrum wrote:
[…] Definitely no turnaround over 7 digits. A small but committed publisher, that is engaged in making PDF documents also accessible for blind people. Hi Steffen,
could you send them this attachment (https://mailman.ntg.nl/archives/list/ntg-context@ntg.nl/message/DS5R3HFLWRFW...) so that they could test whether it works for them? BTW, if this failed, it would be great to have program, version number and OS (to be able to test it further). Many thanks for your help, Pablo

Am 15.02.2025 um 14:16 schrieb Pablo Rodriguez via ntg-context
: On 2/14/25 20:03, Steffen Wolfrum wrote:
[…] Definitely no turnaround over 7 digits. A small but committed publisher, that is engaged in making PDF documents also accessible for blind people. Hi Steffen,
could you send them this attachment (https://mailman.ntg.nl/archives/list/ntg-context@ntg.nl/message/DS5R3HFLWRFW...) so that they could test whether it works for them?
BTW, if this failed, it would be great to have program, version number and OS (to be able to test it further).
Many thanks for your help,
Pablo
Hi Pablo, I already did this some days ago. Of course not the same PDF but probably from a very similar context code. It didn't work. Do I understand you correctly that it world be interesting to know program, version number and OS of the program with which they tested our PDF, unsuccessfully? Regards, Steffen

On 2/15/25 19:29, Steffen Wolfrum wrote:
[...] Do I understand you correctly that it would be interesting to know program, version number and OS of the program with which they tested our PDF, unsuccessfully?
Hi Steffen, on a second thought, only the program would be required. It would be the easiest way to replicate their tests. I mean, if they use a publicly-available validator. Otherwise, it will be really hard to test accessibility for PDF documents by ourselves. But for now, I have tested this with the VeraPDF validator and I have found another issue. If you don’t mind, I will open a new thread. Best, Pablo

\setupbackend[ format={pdf/ua-1}, profile={default_rgb.icc,default_gray.icc}, ] \setuptagging [state=start] \enabledirectives [backend.usetags=mkiv] \starttext a \stoptext is green with VeraPDF! format={pdf/ua-2} is not defined I think, see: https://source.contextgarden.net/tex/context/base/mkiv/lpdf-fmt.lua Greetings Thomas Am 16.02.25 um 10:03 schrieb Pablo Rodriguez via ntg-context:
On 2/15/25 19:29, Steffen Wolfrum wrote:
[...] Do I understand you correctly that it would be interesting to know program, version number and OS of the program with which they tested our PDF, unsuccessfully? Hi Steffen,
on a second thought, only the program would be required.
It would be the easiest way to replicate their tests. I mean, if they use a publicly-available validator.
Otherwise, it will be really hard to test accessibility for PDF documents by ourselves.
But for now, I have tested this with the VeraPDF validator and I have found another issue.
If you don’t mind, I will open a new thread.
Best,
Pablo ___________________________________________________________________________________ If your question is of interest to others as well, please add an entry to the Wiki!
maillist :ntg-context@ntg.nl /https://mailman.ntg.nl/mailman3/lists/ntg-context.ntg.nl webpage :https://www.pragma-ade.nl /https://context.aanhet.net (mirror) archive :https://github.com/contextgarden/context wiki :https://wiki.contextgarden.net ___________________________________________________________________________________

On 2/16/25 11:09, Thomas Meyer wrote:
\setupbackend[ format={pdf/ua-1}, profile={default_rgb.icc,default_gray.icc}, ]
\setuptagging [state=start]
\enabledirectives [backend.usetags=mkiv]
\starttext a \stoptext
is green with VeraPDF!
Many thanks for your testing, Thomas.
format={pdf/ua-2} is not defined I think, see: https:// source.contextgarden.net/tex/context/base/mkiv/lpdf-fmt.lua
It’s defined for MkXL (aka LMTX). See lpdf-fmt.lua from your distribution, since the ConTeXt Garden contains a version which is more than a year old). BTW, the profile mixture has a separate thread. Best, Pablo

Hi Pablo, you are correct: i found lpdf-fmt.lmt. But with format={pdf/ua-2} VeraPDF is still red. Greetings Am 16.02.25 um 11:36 schrieb Pablo Rodriguez via ntg-context:
On 2/16/25 11:09, Thomas Meyer wrote:
\setupbackend[ format={pdf/ua-1}, profile={default_rgb.icc,default_gray.icc}, ]
\setuptagging [state=start]
\enabledirectives [backend.usetags=mkiv]
\starttext a \stoptext
is green with VeraPDF! Many thanks for your testing, Thomas.
format={pdf/ua-2} is not defined I think, see: https:// source.contextgarden.net/tex/context/base/mkiv/lpdf-fmt.lua It’s defined for MkXL (aka LMTX). See lpdf-fmt.lua from your distribution, since the ConTeXt Garden contains a version which is more than a year old).
BTW, the profile mixture has a separate thread.
Best,
Pablo ___________________________________________________________________________________ If your question is of interest to others as well, please add an entry to the Wiki!
maillist :ntg-context@ntg.nl /https://mailman.ntg.nl/mailman3/lists/ntg-context.ntg.nl webpage :https://www.pragma-ade.nl /https://context.aanhet.net (mirror) archive :https://github.com/contextgarden/context wiki :https://wiki.contextgarden.net ___________________________________________________________________________________

On 2/16/25 11:57, Thomas Meyer wrote:
Hi Pablo,
you are correct: i found lpdf-fmt.lmt.
But with format={pdf/ua-2} VeraPDF is still red.
Yes, the main problem I found was the reason I started a new thread (https://mailman.ntg.nl/archives/list/ntg-context@ntg.nl/message/4IIHP4EGYVY3...). Best, Pablo

Oh, sorry! Am 16.02.25 um 12:02 schrieb Pablo Rodriguez via ntg-context:
Yes, the main problem I found was the reason I started a new thread (https://mailman.ntg.nl/archives/list/ntg-context@ntg.nl/message/4IIHP4EGYVY3...).
Best,
Pablo ___________________________________________________________________________________ If your question is of interest to others as well, please add an entry to the Wiki!
maillist : ntg-context@ntg.nl / https://mailman.ntg.nl/mailman3/lists/ntg-context.ntg.nl webpage : https://www.pragma-ade.nl / https://context.aanhet.net (mirror) archive : https://github.com/contextgarden/context wiki : https://wiki.contextgarden.net ___________________________________________________________________________________

Hi Pablo, they tested the test file: they could not see alt-text: Acrobat Pro 24.005 / Windows (x86) Steffen
Am 16.02.2025 um 10:03 schrieb Pablo Rodriguez via ntg-context
: On 2/15/25 19:29, Steffen Wolfrum wrote:
[...] Do I understand you correctly that it would be interesting to know program, version number and OS of the program with which they tested our PDF, unsuccessfully?
Hi Steffen,
on a second thought, only the program would be required.
It would be the easiest way to replicate their tests. I mean, if they use a publicly-available validator.
Otherwise, it will be really hard to test accessibility for PDF documents by ourselves.
But for now, I have tested this with the VeraPDF validator and I have found another issue.
If you don’t mind, I will open a new thread.
Best,
Pablo ___________________________________________________________________________________ If your question is of interest to others as well, please add an entry to the Wiki!
maillist : ntg-context@ntg.nl / https://mailman.ntg.nl/mailman3/lists/ntg-context.ntg.nl webpage : https://www.pragma-ade.nl / https://context.aanhet.net (mirror) archive : https://github.com/contextgarden/context wiki : https://wiki.contextgarden.net ___________________________________________________________________________________

On 2/18/25 13:02, Steffen Wolfrum wrote:
Hi Pablo,
they tested the test file: they could not see alt-text: Acrobat Pro 24.005 / Windows (x86).
Sorry, Steffen, I need to investigate this further. Although I must confess that if Acrobat Pro (not Reader) is the only way to test this, I don’t have access to it. In any case, I’m afraid testing will take time, because I have to dedicate myself to other tasks. Best, Pablo

Am 18.02.2025 um 17:19 schrieb Pablo Rodriguez via ntg-context
: On 2/18/25 13:02, Steffen Wolfrum wrote:
Hi Pablo,
they tested the test file: they could not see alt-text: Acrobat Pro 24.005 / Windows (x86).
Sorry, Steffen, I need to investigate this further.
Although I must confess that if Acrobat Pro (not Reader) is the only way to test this, I don’t have access to it.
In any case, I’m afraid testing will take time, because I have to dedicate myself to other tasks.
Hi Pablo, I’d be happy if I could support your investigations. At least I will communicate and forward any test file to those who have Acrobat Pro. Yours, Steffen

Pablo, the last Acrobat Pro without subscription is version from 2020. On a macOS I tested your test file with this version and the result was surprising, see screenshots via wetransfer: https://we.tl/t-eW1zqTncZw https://we.tl/t-eW1zqTncZw Steffen
Am 19.02.2025 um 07:22 schrieb Steffen Wolfrum
mailto:context@st.estfiles.de>: Am 18.02.2025 um 17:19 schrieb Pablo Rodriguez via ntg-context
mailto:ntg-context@ntg.nl>: On 2/18/25 13:02, Steffen Wolfrum wrote:
Hi Pablo,
they tested the test file: they could not see alt-text: Acrobat Pro 24.005 / Windows (x86).
Sorry, Steffen, I need to investigate this further.
Although I must confess that if Acrobat Pro (not Reader) is the only way to test this, I don’t have access to it.
In any case, I’m afraid testing will take time, because I have to dedicate myself to other tasks.

On 2/21/25 12:32, Steffen Wolfrum wrote:
Pablo,
the last Acrobat Pro without subscription is version from 2020. On a macOS I tested your test file with this version and the result was surprising, see screenshots via wetransfer: https://we.tl/t-eW1zqTncZw.
Hi Steffen, the relevant info is missing in the frame I added to your first screenshot. Just as a comment, pngquant does a great job reducing PNG screenshots (most of the time). Starting a new thread about the non-standard tag. Many thanks for your patience, Pablo

Am 25.02.2025 um 17:01 schrieb Pablo Rodriguez via ntg-context
mailto:ntg-context@ntg.nl>: On 2/21/25 12:32, Steffen Wolfrum wrote:
Pablo,
the last Acrobat Pro without subscription is version from 2020. On a macOS I tested your test file with this version and the result was surprising, see screenshots via wetransfer: https://we.tl/t-eW1zqTncZw https://we.tl/t-eW1zqTncZw.
Hi Steffen,
the relevant info is missing in the frame I added to your first screenshot.
Just as a comment, pngquant does a great job reducing PNG screenshots (most of the time).
Starting a new thread about the non-standard tag.
Hi Pablo, sorry for missing parts. Here attached you may find these info. Thanks for mentioning pngquant, works great (as we can see here)! (I hope this is still the right thread?) Yours, Steffen
participants (5)
-
Hans Hagen
-
Hans Hagen
-
Pablo Rodriguez
-
Steffen Wolfrum
-
Thomas Meyer