- From: Yehuda Katz <wycats@gmail.com>
- Date: Tue, 25 Feb 2014 11:42:21 -0800
- To: Andrea Giammarchi <andrea.giammarchi@gmail.com>
- Cc: Brian Kardell <bkardell@gmail.com>, Robin Berjon <robin@w3.org>, "public-nextweb@w3.org" <public-nextweb@w3.org>
- Message-ID: <CAMFeDTWW7ZTRFVjC6qdKe_7GW15XKn-vH7v6eTrEtW6p85dnBw@mail.gmail.com>
I think this list of issues is actually good evidence that we should build from the ground up here and let the library ecosystem drive this for a bit before getting the feature into the platform. I'd expect the entire process to take no longer than trying to standardize it a priori, and be less error prone. Yehuda Katz (ph) 718.877.1325 On Tue, Feb 25, 2014 at 11:40 AM, Andrea Giammarchi < andrea.giammarchi@gmail.com> wrote: > my quick take on those questions: > > ISSUE 1: JSON root types > every form field has a name and since quite ever. I've never seen <input > name="[]"/> so that the server will receive a list instead of an object so > I don't see why we would go in any other direction. Object is good > > ISSUE 2: Path syntax > if that works out of the box in Rails and needs few adjustments in PHP > and/or Java I would say it's already good. > It looks also very "natural" and JSON friendly so I think path as > described is OK > > ISSUE 3: Base64 > FormData accepts blobs, which is the only thing that would make it more > suitable for files. > This is true only if there won't be the possibility to send > gzipped/deflated data from the client though, as it's demonstrated base64 > won't blow "that much" compared with binary and it also will play very well > with big forms and repeated properties. > Latter is true in any case for every big form we know or use already but > here, going base64, I'd like to see such possibility implemented! > > ISSUE 4: Is append needed? > append is very used in PHP land and I think are quite handy but I also > remember these are not widely adopted. > I'd chose them if common adoption is higher than non adoption (like a > passive vote from the web community) > > ISSUE 5: Boolean radio/checkbox > Usually radio and checkboxes are sent only if checked so that server-side > checks the presence, not its value, right? not an issue, IMO > > ISSUE 6: Ignore charsets > if charset can solve anything it should be probably supported with an > UTF-8 as default > > > On Tue, Feb 25, 2014 at 11:19 AM, Brian Kardell <bkardell@gmail.com>wrote: > >> >> >> >> On Tue, Feb 25, 2014 at 11:02 AM, Robin Berjon <robin@w3.org> wrote: >> >>> Hi, >>> >>> I put together a small spec to enable having HTML form data submitted >>> directly as JSON, with structure. I'd be curious to hear your feedback: >>> >>> http://darobin.github.io/formic/specs/json/ >>> >>> Thanks! >>> >>> -- >>> Robin Berjon - http://berjon.com/ - @robinberjon >>> >>> >> >> Robin, >> >> Some thoughts all over the place... >> >> Would this also modify the FormData serialization in XMLHttpRequest[1]. >> That would be awesomely more useful to me. It would also be wickedly >> handy if there were a way to get at it (ie, some DOM method), probably more >> useful than even the other two since it would let me polyfill the other two >> with 1 underlying implementation that explains it. <form>.toObjectModel() >> or something which provides a single canonical non-DOM object >> representation of the form which could be serialized and deserialized could >> be used to explain how enctype="application/json" and corresponding >> FormData in ajax would work too. Of course, this begs the question of >> whether there is a corresponding <form>.fromObjectModel() which could >> populate the form given some data (kinda awesome), and whether both of >> those are maybe just natively handleable custom serializations ala >> JSON.stringify(form) and JSON.parse(form).. >> >> >> http://www.w3.org/TR/XMLHttpRequest2/#interface-formdata >> >> >> >> -- >> Brian Kardell :: @briankardell :: hitchjs.com >> > >
Received on Tuesday, 25 February 2014 19:43:10 UTC