It's powered by a mega-corp AI, it has a Liquid Mode, but it's not a T-1000. It's Adobe's PDF auto-reflow for mobile
- Reference: 1600934226
- News link: https://www.theregister.co.uk/2020/09/24/adobe_liquid_mode_pdf_mobile/
- Source link:
Adobe’s Portable Document Format (PDF) dates from 1993, based on the PostScript language used to describe documents accurately to printing. A PDF document should as far as possible look the same on-screen and when printed, down to the last hyphen and the exact position of a caption under an image.
This makes it ideal for proof-reading documents before printing, and for digital documents where preserving layout is important. The ability to view PDFs on any computer, and the fact that it is primarily a read-only format, won PDF wide use for document transmission beyond the print industry, and Adobe built on its success with sophisticated document management and workflow systems.
The software giant says that there are more than 2.5 trillion PDFs out there, and that the demand for digital everything, stimulated perhaps by the coronavirus pandemic lockdown, has increased use of its PDF-based services such as Adobe Sign. But there's a problem: it's not great with multi-screen digital environments, where documents should adapt according to the size and resolution of the display.
On the web we now expect a responsive layout, but PDFs remain stubbornly fixed. Tapping or double-clicking a PDF all too often renders it too small to read, and when zoomed to a readable size horizontal scrolling may be necessary.
Earlier this month usability veteran Jakon Nielsen [1]described PDF as “still unfit for human consumption, 20 years later.” Explaining, he said “PDFs are typically large masses of text and images. The format is intended and optimized for print. It’s inherently inaccessible, unpleasant to read, and cumbersome to navigate online. Neither time nor changes in user behavior have softened our evidence-based stance on this subject.”
Teach me, Sensei
As if it had read his mind, Adobe has now introduced a new mode for its Acrobat Reader which is intended to overcome these exact issues. Called Liquid Mode, the service uses AI – Adobe’s Sensei service – to reformat the document – the artificial intelligence being called upon to parse the structure of a document, what is a heading, what is body text and so on. Liquid Mode is now available in Reader for iOS and Android, and is promised for desktop Reader and for web browsers (though many browsers have their own built-in PDF readers).
The effect is that the document automatically becomes responsive, reflowing to fit the screen, and with options to control the font size and line spacing. Pinch to zoom no longer works, but instead users can tap Settings according to preference, with no need for horizontal scrolling.
Adobe Illustrator's open source rival Inkscape delivers v1.0.1 - with experimental Scribus PDF export [2]READ MORE
Adobe says Liquid Mode is but the beginning of a [3]new PDF era “Imagine using AI to replace months of intensive document research with a quick query that cross-references millions of PDFs in your organization at once,” it says, seemingly on a similar path to Microsoft which unveiled SharePoint Syntex at Ignite, also for intelligent document categorisation with AI.
Adobe promises to bring “more Adobe Sensei technology into all of our document products and services.” With regard to PDF, Liquid Mode is the start of an effort to give people tools “to view and extract content from PDFs, without changing the PDF itself.”
We took a quick look at Liquid Mode on iOS. The good news is that it works as advertised, on documents that it supports. There are some downsides though.
First, to enable Liquid Mode users have to sign in to Adobe’s cloud. Liquid Mode sends content to Adobe in order to process it, raising the usual concerns about privacy, particularly since Adobe is also a big player in digital marketing and personalisation - though there is no suggestion that it is using Liquid Mode content for this purpose and the company says, “files are processed in our secure data servers and immediately deleted after the experience is generated.”
We soon discovered some limitations. Presented with a complex PDF containing multiple tables, Reader declared that “Liquid Mode isn’t yet available for this file” and pointed us to [4]this document which has a more sober assessment of the feature than the company’s breathless press release.
Files cannot use Liquid Mode if they are more than 200 pages long, more than 10MB, scanned, converted from presentations, or contain “complex layouts” such as tables and charts, invoices, statements and receipts, forms, or “non-standard page sizes”.
Sensei, it seems, prefers more straightforward documents. Understandable, but it means that the pain of PDF, especially on mobile, is not altogether over yet. ®
Get our [5]Tech Resources
[1] https://www.nngroup.com/articles/pdf-unfit-for-human-consumption/
[2] https://www.theregister.com/2020/09/09/inkscape_1_0_1/
[3] https://blog.adobe.com/en/2020/09/23/adobe-unveils-ambitious-multi-year-vision-for-pdf-introduces-liquid-mode.html
[4] https://docs.adobe.com/content/help/en/document-cloud-mobile/using/liquidmode-en.html
[5] https://whitepapers.theregister.com/
Re: RE: Ideal to proof read before printing
> No-one needs this
Yes we do. We, the people who on occasion want to read a PDF document whilst on the go, have no choice over the format we find it in. Remember - a huge number of instruction manuals are most easily found in PDF format.
Your points about publishers choosing more suitable formats is valid, but that doesn't address the fact that there are millions of existing PDFs on the web, some of which might just have the information I need to fix something.
Should we need an Adobe account to read PDFs on a small screen in a sane fashion? No, of course not. But some sort of text reflow - or even just more intelligent zooming and scrolling for common two-column PDFs - is very useful indeed.
It is that for that reason that use the Gaaiho PDF reader on my Adroid phone.
( I haven't tested Gaaiho Reader extensively, but I was driven to try it out after being frustrated by the very issues Adobe are only now claiming to address).
Re: RE: Ideal to proof read before printing
I think the point is that it *cannot* work for a lot of PDFs.
PDF is a wrapper format - the contents can be a sequence of a load of words/letters, in which case liquid mode has a chance of working, or it can be an image of a page, in which case it will not work.
The clue about whether it will work or not on any given document would be to try and copy an line of text out of the document: if you can copy the text and paste it into notepad, it'll possibly work in liquid mode.
I don't see how you can proofread something for printing in a format that doesn't exactly represent the printed page that you expect to get at the end of the process. Or has proofreading been reduced to "the author got all the words in the right order"?
Re: RE: Ideal to proof read before printing
From the description -
"Files cannot use Liquid Mode if they are more than 200 pages long, more than 10MB, scanned, converted from presentations, or contain “complex layouts” such as tables and charts, invoices, statements and receipts, forms, or “non-standard page sizes”."
it sounds like it is just OCR smacked around with the "AI" BrandoStick, with the result formatted to fit on the requested screen.
Re: RE: Ideal to proof read before printing
To be clear:
There is a need for a PDF reader to make the reading of existing PDFs easier on mobile screens.
I did not mean that there was a need for Adobe's specific solution.
I did mean that that there is a clear problem: reading existing PDFs with the default reader on Android can be a pain. This problem has been partially addressed by PDF readers from 3rd parties that offer text reflow and more zoom and scrolling options.
Even the above suggestion that PDF text be copied and pasted into a text editor is just more evidence that the problem exists.
Mage's argument that there is no need for a better PDF reader of any description because PDFs shouldn't be used isn't helpful in the context of, for example, consulting an instruction manual PDF for a washing machine error code. Mage's 'solution' does nothing to make for the millions of extant PDF files on the web easier to read when one doesn't have a big screen to hand.
Unfortunately it still won't fix the all too common completely moronic use of PDF documents such as:
Where everything is output as images, no text at all (complete usability and accessibility failure)
The document character order is different to thevisual character display order (another complete accessiblity failure, but just about usable if crippled in use)
Where the document is rendered for a specific orientation, such as rendering two pages side by side to allow the multi page document to be "easier to read on the web" (accessible, but largely unusable)
I've seen a couple of occasions where a pdf full of scans made some sense.
One was preserving a manual for abandonware that we were still using. Naturally, I fed that to OCR software, but correcting it afterwards was time consuming.
The other was for archiving against FOI requests. When I laughed at that it was explained to me that any hand written notes needed to be captured too.
Hmm, maybe I need two more exceptions to my LARTing comment below.
> moronic use of PDF documents such as: Where everything is output as images, no text at all (complete usability and accessibility failure)
No text is handy when sending, for example, a poster created in Illustrator to a printing agency in PDF format. Converting the text to vector outlines means that you don't have to worry about the printing agency having all the fonts you have used. Obviously in this use case the lack of edibility is not an issue - the creator retains the original document and doesn't want the printer to change anything.
Just unfit for the purpose
Basically, anyone who writes pdf for any use case other than print or proofing needs to be re-educated with the biggest LART available. There are multiple formats designed to reflow as needed, although many alleged designers insist on nobbling that functionality in the name of 'pixel perfection.'
It's been nearly 30 years of the web. It's long past time we dropped what Ted Nelson mocks as, 'paper under glass.'
Re: Just unfit for the purpose
What often happens is that a PDF will be created so that an instruction manual can be printed. End user loses instruction manual. Years later user needs to consult manual. They find it easier to find PDF version on web than to find original hard copy.
Re: Just unfit for the purpose
Fuck Jakon Nielsen.
PDFs are perfect for things like instruction manuals, where things need to be laid out properly and clearly, without web-style stupid shit like the image on one page and the caption on another.
Re: Just unfit for the purpose
What is this 'page' of which you speak? That's a squashed tree concept isn't it?
Re: Just unfit for the purpose
BTW, sensible use of a block level element will group the picture and caption together. Semantic markup and all that.
But is it Flash?
Could it be a Flash viewer for pdf?
RE: Ideal to proof read before printing
Not if it's reflowing on a small screen.
The ENTIRE point of PDF is to proof for Printing. Or Print. You need a big screen.
We have two superior techs to read documents on small screens: The epub based ebook (a Kindle AZW/KF8 format is basically the same) or HTML.
No-one needs this and it won't work for many existing PDFs. It will encourage idiots to use PDFs when they ought to produce ebooks (PDFs can be books in electronic form, they are NOT eBooks) or HTML. The only sensible use cases for PDFs are to proof something for print (so it CAN NOT be reflowed!) or to distribute for print or as a wrapper for a scanned multipage document, superior to multipage TIF, that can optionally have OCR for search. PDF forms are a stupid evil invention.