4 ms·

No. Terrible idea. The authors take the perspective that content should be easy for browsers to render. That's important but shouldn't be the only consideration

No. Terrible idea. The authors take the perspective that content should be easy for browsers to render. That's important but shouldn't be the only consideration. Browsers can render HTML well, but browsers have incredibly complicated rendering engines. This makes HTML non-portable and non-exportable, not printable, not sharable. I'm sure for each one of those you can create a workaround to bridge the gap but the big picture is that HTML content tends to be dynamic and rendered by a heavy server component from database which makes creating a true standard actually quite hard. HTML actually really sucks as a portable document standard without some extra semantic definition and a container format. So to make HTML workable you have to do something like what docx and odf do: XML definition+resources in zip container ... which now look a lot like PDF. >We also intend to build functionality for users to automatically generate accessible PDFs from HTML documents. How about the other way.. PDFs -> HTML documents.

anticensor 7y ago
That would not work perfectly as some PDFs do not have the text stored as "text".
- macspoofing 7y ago
  That's not the fault of PDF. You can have an HTML document with image links as well. The problem is that you have no standard way of exporting and storing this kind of document.
qqii 7y ago
> We also intend to build functionality for users to automatically generate accessible PDFs from HTML documents. This would mean that publishers will only need to create and maintain one document, but users will still be able to download a PDF if they need to. Oh I see you've added that. The problem with PDF to HTML is thst your trying to fit something static back into something dynamic. I also don't see how pdf rendering is any less complicated. There are many browser implementations that can easily render gov.uk without issues.
- macspoofing 7y ago
  This raises the question of how they create these 'HTML documents'. As far as I'm aware, there really isn't an HTML document format. More than likely, these documents will be created by some web-based system (maybe even proprietary and cloud-based, making portability extremely hard) and simply exported as PDF later. That's how every web-based content management system does it today. It's very convenient right now, but when it comes to document maintenance for the next few decades, I'm not sure that's better than something like PDF or ODF.
  - petepete 7y ago
    The code is available on GitHub, you can see for yourself. It's mostly written in a customised markdown variant called Govspeak[0]. You appear to be jumping to ridiculous conclusions. Decisions on how we deliver content are made on the back of a huge amount of research. It's more accessible than it's ever been, other governments and organisations are following our lead. To suggest PDF, ODF or other crippled, inaccessible, OS-constrained formats over HTML - something everyone can read with any of their devices - shows you need to do more research before posting your ill-informed views. [0] https://github.com/alphagov/govspeak https://github.com/alphagov/govspeak Edit: Also, web page have plenty of other advantages. Indexed by every search engine, bookmarkable/shareable headings. What other medium offers that?
  - qqii 7y ago
    You can find out more about how GDS operates on their github[0]. [0]: https://alphagov.github.io/ https://alphagov.github.io/
projektfu 7y ago
You could offer content in an ebook format (ePub, mobipocket) which is just HTML and resources with styles and offer it through the web as straight HTML with the option of downloading the container.
- macspoofing 7y ago
  For sure. But notice you need to define another document standard because you probably don't want to just use an ebook format since that won't capture all your use-cases. So why not just use PDF, docx, or ODF?
  - projektfu 7y ago
    Honestly because those formats are more restrictive (in an engineering sense), proprietary (in the sense of who really controls them) and difficult to parse. You’d need some sort of special subset of the format to make it so that people’s use didn’t obscure information, as is easily doable with scrambled PDFs and the like. You’d have to validate that format and reject submissions that didn’t use it. Laziness would probably lead to whatever the printer driver produces being the de facto standard.
mwcampbell 7y ago
> you have to do something like what docx and odf do: XML definition+resources in zip container ... which now look a lot like PDF. Only superficially. .docx and .odt are single files as PDF is, but the contents of those files are very different. A PDF document is designed around the assumption of printed pages with a fixed layout. Yes, tagged PDF has been retrofitted onto the base PDF format for accessibility, but most PDFs that I've seen in the wild do not use this extra feature. To understand why starting with a PDF and converting to HTML is a bad idea, try any PDF-to-HTML converter you can find; there's the pdftohtml utility included with Xpdf and Poppler, but any of the commercial alternatives are not a lot better. Once you learn more about how PDF works, you'll understand why PDF to HTML conversion can't be very good.
shakna 7y ago
> HTML actually really sucks as a portable document standard without some extra semantic definition and a container format. You're probably looking for the WARC [0] ISO standard. Widely supported and used. Including by the UK [1]. WARC is simple enough that tools to create or extract from it are dime a dozen. As for the incredibly complicated rendering engines to look at the extracted HTML when you want, then even links doesn't struggle to render the content. [0] https://www.loc.gov/preservation/digital/formats/fdd/fdd000236.shtml https://www.loc.gov/preservation/digital/formats/fdd/fdd0002... [1] https://www.nationalarchives.gov.uk/pronom/fmt/289 https://www.nationalarchives.gov.uk/pronom/fmt/289

anticensor 7y ago

That would not work perfectly as some PDFs do not have the text stored as "text".

macspoofing 7y ago
That's not the fault of PDF. You can have an HTML document with image links as well. The problem is that you have no standard way of exporting and storing this kind of document.

qqii 7y ago

> We also intend to build functionality for users to automatically generate accessible PDFs from HTML documents. This would mean that publishers will only need to create and maintain one document, but users will still be able to download a PDF if they need to. Oh I see you've added that. The problem with PDF to HTML is thst your trying to fit something static back into something dynamic. I also don't see how pdf rendering is any less complicated. There are many browser implementations that can easily render gov.uk without issues.

macspoofing 7y ago
This raises the question of how they create these 'HTML documents'. As far as I'm aware, there really isn't an HTML document format. More than likely, these documents will be created by some web-based system (maybe even proprietary and cloud-based, making portability extremely hard) and simply exported as PDF later. That's how every web-based content management system does it today. It's very convenient right now, but when it comes to document maintenance for the next few decades, I'm not sure that's better than something like PDF or ODF.
- petepete 7y ago
  The code is available on GitHub, you can see for yourself. It's mostly written in a customised markdown variant called Govspeak[0]. You appear to be jumping to ridiculous conclusions. Decisions on how we deliver content are made on the back of a huge amount of research. It's more accessible than it's ever been, other governments and organisations are following our lead. To suggest PDF, ODF or other crippled, inaccessible, OS-constrained formats over HTML - something everyone can read with any of their devices - shows you need to do more research before posting your ill-informed views. [0] https://github.com/alphagov/govspeak https://github.com/alphagov/govspeak Edit: Also, web page have plenty of other advantages. Indexed by every search engine, bookmarkable/shareable headings. What other medium offers that?
- qqii 7y ago
  You can find out more about how GDS operates on their github[0]. [0]: https://alphagov.github.io/ https://alphagov.github.io/

projektfu 7y ago

You could offer content in an ebook format (ePub, mobipocket) which is just HTML and resources with styles and offer it through the web as straight HTML with the option of downloading the container.

macspoofing 7y ago
For sure. But notice you need to define another document standard because you probably don't want to just use an ebook format since that won't capture all your use-cases. So why not just use PDF, docx, or ODF?
- projektfu 7y ago
  Honestly because those formats are more restrictive (in an engineering sense), proprietary (in the sense of who really controls them) and difficult to parse. You’d need some sort of special subset of the format to make it so that people’s use didn’t obscure information, as is easily doable with scrambled PDFs and the like. You’d have to validate that format and reject submissions that didn’t use it. Laziness would probably lead to whatever the printer driver produces being the de facto standard.

mwcampbell 7y ago

> you have to do something like what docx and odf do: XML definition+resources in zip container ... which now look a lot like PDF. Only superficially. .docx and .odt are single files as PDF is, but the contents of those files are very different. A PDF document is designed around the assumption of printed pages with a fixed layout. Yes, tagged PDF has been retrofitted onto the base PDF format for accessibility, but most PDFs that I've seen in the wild do not use this extra feature. To understand why starting with a PDF and converting to HTML is a bad idea, try any PDF-to-HTML converter you can find; there's the pdftohtml utility included with Xpdf and Poppler, but any of the commercial alternatives are not a lot better. Once you learn more about how PDF works, you'll understand why PDF to HTML conversion can't be very good.

shakna 7y ago

> HTML actually really sucks as a portable document standard without some extra semantic definition and a container format. You're probably looking for the WARC [0] ISO standard. Widely supported and used. Including by the UK [1]. WARC is simple enough that tools to create or extract from it are dime a dozen. As for the incredibly complicated rendering engines to look at the extracted HTML when you want, then even links doesn't struggle to render the content. [0] https://www.loc.gov/preservation/digital/formats/fdd/fdd000236.shtml https://www.loc.gov/preservation/digital/formats/fdd/fdd0002... [1] https://www.nationalarchives.gov.uk/pronom/fmt/289 https://www.nationalarchives.gov.uk/pronom/fmt/289