Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Yes, maybe generally; but let’s talk about the specific case here — filling of complex PDF forms.

When a PDF that has interactive form fields, calculated auto-populated fields, fields that are enabled/disabled according to the inputs of other fields, etc. — the organization that created it (usually government or education) usually does that because they want you to fill it out using a PDF viewer; save it (which will persist the form inputs “into” the resulting PDF); and then submit the modified PDF file back to them. They want this, because they can use automated backend processes to extract the data from the PDF. They don’t want you to just print out the thing and fill it out. In fact, many such “fillable” PDFs start off in a state with many of their form-fields disabled and voided, such that printing them out in that state would result in a form you can’t really write on!

So, at that point, why didn’t they just make the PDF a web page? They’ve essentially reinvented a web form, but with extra steps. The only benefit a client gets is the ability to edit and save the form offline (but that can be done in a browser, too, with local storage); and furthermore, the ability to treat the resulting filled form as a file, moving it around before you submit it. But the cases where you need that are very niche, compared to the cases where you can just direct employees to your Intranet portal.



1. A webpage form requires a server to be up and running, which requires an IT person to manage it, separate from the dept making the form. PDF forms can be created by a person given the right tools (I think Word does it)

2. IT person + webserver costs have to included in the budget somewhere. Which can be a big problem.

3. The webpage form can fail, and the support for it has to be provided by the IT dept. If the PDF form fails, dept can handle it on its own, and will often accept a filled+scanned print out of the PDF form.

4. Adding to the point above, PDF forms degrade gracefully, If they don't work, or internet doesn't work, or someone is on holiday, you can still print, fill and hand them in person. Webpages can degrade catastrophically where you whole dept grinds to halt while the IT person tries to fix the problem.


Re: all four of your points — see my sibling post. I'm not talking about encapsulated-PostScript "Print and Fill" forms (which do certainly degrade gracefully), or even open-standard PDF "Fill and Print" forms (which degrade gracefully if you don't set them up with a bad default state where there's big "N/A" text over all the disabled fields until you fill in other fields.)

Instead, I'm talking about the PDFs you can basically only load in Acrobat (though, other PDF viewers do try to render them, to varying success) that actually do data-binding to some remote database; do XHRs to submit the form data on success; do "online" onBlur-XHR-esque field validation; generate new output PDFs using scripting, from scratch when you ask them to save/print; etc.

These are applications, not documents. You can't print them. You just use Acrobat as a glorified application host to fill and submit them. (You can press Ctrl+P to get Acrobat to request to the loaded PDF application that it perform some scripted action to generate a print output. This may or may not do anything, depending on how the PDF was created. It usually just pops a "Printing is not implemented for this form" box. It certainly won't work in non-Acrobat PDF viewers.)

When other PDF viewers say they don't support "fillable PDF eForms", these are the things they're talking about. They usually support "Fill and Print" forms just fine, because "Fill and Print" forms are a somewhat-sane format, rather than being a competitor to Lotus Notes.


I understand better what you are saying. I don't think I have ever seen any PDF forms that require an internet connection. The Canadian Visa application forms have inbuilt validation code, that checks the form, and once you upload it, I believe data is extracted into a database.

The benefit of these forms is that the validated form that you submit online is actually printable. Which means that what you see on your screen/paper is pixel by pixel identical to what Canada receives, and therefore _legally_, there is no confusion about what was communicated between Canada and the candidate.

Webforms are not as strongly accepted as such by courts. Because they have to be manipulated further before being printed.

I have read a bunch of your replies, and you are thinking of all the technical reasons why webforms are better than PDF (you are right in that), but PDFs have legal and operational and budgetary advantages, that are more relevant to various organizations.


> In fact, many such “fillable” PDFs start off in a state with many of their form-fields disabled and voided, such that printing them out in that state would result in a form you can’t really write on!

I have never seen this. Do you have an example? Every use if fillable PDFs I have encountered is a use case where submitting a handwritten form is still an option.

> The only benefit a client gets is the ability to edit and save the form offline (but that can be done in a browser, too, with local storage); and furthermore, the ability to treat the resulting filled form as a file, moving it around before you submit it.

I have yet to see a web form that actually saves a readable, properly-formatted, self-contained, easy to access, fully-offline copy.

> But the cases where you need that are very niche, compared to the cases where you can just direct employees to your Intranet portal.

This is not a trivial need; most forms sent as fillable PDFs need to or should be retained for some period after submission. Also, I don't know what "employees" and "Intranet" has to do with anything.

You are also missing the use case where a form legally requires a live signature from one or more parties and need to be printed, even if just to scan and return. I recently had to do this for some insurance paperwork.


> You are also missing the use case where a form legally requires a live signature from one or more parties and need to be printed, even if just to scan and return. I recently had to do this for some insurance paperwork.

My company have to do this for one state government. They required the signature to be written black inked. It is PITA to do since we all have digital signature set up. But nope, this state government required the written signature.


The Canadian visa application form is an example.


> I have never seen this. Do you have an example?

I don't have one on-hand, no. But I've certainly had to fill them out in the past. IIRC an especially-bad one came in the form [heh] of a student-loan application for the college I attended. It was essentially a Hypercard stack in the guise of a PDF.

Here are some early Adobe marketing materials (as a PDF, because of course it is) talking about the advantages of "eForm Solutions": https://planetpdf.com/planetpdf/pdfs/pdf2k/02E/ldefurio_pdff...

It sounds like every PDF form you've ever dealt with is what Adobe, in this brochure, calls a "Type 1: Print and Fill" or "Type 2: Fill and Print" form. But Type 3 and Type 4 forms do exist in the wild! (They're not often created any more; most of the ones that exist now are from around a decade or two ago, when Adobe was really pushing this idea.) Creating such forms was basically the point of Acrobat as a software product.

When PDF viewers (e.g. Apple Preview) say they don't support "PDF forms", they're not talking about Type 2 forms. They usually support those just fine. They're talking about Type 3 and Type 4 forms. And more specifically, the ones that use Adobe's proprietary AcroForms data-embedding system, rather than the open-standard XFA data-embedding system.

(I could swear I saw an HN post about the horrors of AcroForms once, but I can't find it now.)

> I have yet to see a web form that actually saves a readable, properly-formatted, self-contained, easy to access, fully-offline copy.

To be clear, that was what I meant by the second qualifier, "as a file." Browsers support persisting the state of the form. Just, not as a file. They persist the state internally, when the form's author does the client-side Javascript work to enable that.

For the use-case where the user wants to stop filling out the form for now (e.g. because they don't have some required information on-hand), and then come back to it to finish it later, in-browser persistence works perfectly well.

Even cleaner, though, is just building a web-form as a wizard, where fields are submitted one-at-a-time, and you can also freely navigate to previously-filled "steps" to change your answers. That doesn't even require JavaScript; just pure 90s HTML-generated-on-the-backend. Most government sites that thought PDF eForms were a good idea, are now falling back to this approach.

> Also, I don't know what "employees" and "Intranet" has to do with anything.

Secure installations. The main use-case for fillable PDFs (as can be seen in Adobe's marketing brochure, where "government" is the core client) is a case where public or cloud solutions just aren't tenable, i.e. in secure government/military/etc. installations, where the workstations are air-gapped from the public Internet. In such a case, PDF forms can still be sent around via a local non-Internet-routable email server, for the workers there to fill in.

Today, this need can be served just as well by setting up a non-Internet-routable web portal for those same workers to use. But back in the 90s and 00s, "Intranet web portals" were a fancy thing only the most forward of IT bigcorps had on offer. They had Intranets, for sure, but they weren't hosting web-apps on them.

So, what did they do instead? Well, Adobe had two main competitors in the "eForm" market:

• Lotus Notes form documents, connecting to a Lotus Domino database server;

• Microsoft Excel sheets that use VBA to data-bind to an accessible Microsoft Access database file sitting on an SMB network share.

None of these "forms" were hand-submittable. They're all little self-contained interactive applications, that happen to look like forms.

AcroForms did have the fancy property, though, that the AcroForms application-PDF could generate or export a bog-standard output-PDF representing the filled form. But that's not actually a modified copy of the source PDF. That's the PDF using scripting to generate you another PDF, from scratch.

------

To be clear, I agree with all the stuff you're talking about; those are all valid use-cases for "PDFs" (i.e. encapsulated PostScript containers.) But they're not what I mean by "PDF forms." I mean the Type 3/4 forms referred to above. There's no reason, in the modern era, that one would implement one of these Type 3/4 "eForm solutions", instead of just putting up a webpage.

If you need an e-signature at the end, have them fill out the web form, then generate a raw PostScript PDF representing their inputs, and let them sign it by dropping a signature vector image on the dotted line in any standard PDF viewer.


The use case you're describing wasn't feasible until about 20 years after PDFs were introduced. Web Storage isn't that old, has only recently become widely deployed, and in a lot of cases is disabled for security concerns.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: