2004/id/draft-ietf-html-fileupload-01.txt


INTERNET-DRAFT                                                 E. Nebel
Form-based File Upload in HTML                              L. Masinter
draft-ietf-html-fileupload-01.txt                     Xerox Corporation
Expires May 30, 1995                                  November 30, 1994


                Form-based File Upload in HTML

Status of this Memo

  This document is an Internet-Draft.  Internet-Drafts are working
  documents of the Internet Engineering Task Force (IETF), its areas,
  and its working groups.  Note that other groups may also distribute
  working documents as Internet-Drafts.

  Internet-Drafts are draft documents valid for a maximum of six months
  and may be updated, replaced, or obsoleted by other documents at any
  time.  It is inappropriate to use Internet-Drafts as reference
  material or to cite them other than as ``work in progress.''

  To learn the current status of any Internet-Draft, please check the
  ``1id-abstracts.txt'' listing contained in the Internet-Drafts Shadow
  Directories on ftp.is.co.za (Africa), nic.nordu.net (Europe),
  munnari.oz.au (Pacific Rim), ds.internic.net (US East Coast), or
  ftp.isi.edu (US West Coast).


1. Abstract

  Currently, HTML forms allow the producer of the form to request
  information from the user reading the form.  These forms have proven
  useful in a wide variety of applications in which input from the
  user is necessary.  However, this capability is limited because HTML
  forms don't provide a way to ask the user to submit files of data.
  Service providers who need to get files from the user have had to
  implement custom user applications.  (Examples of these custom
  browsers have appeared on the www-talk mailing list.)  Since
  file-upload is a feature that will benefit many applications, this
  draft proposes an extension to HTML to allow information providers
  to express file upload requests uniformly, and a MIME compatible
  representation for file upload responses.  This draft also includes
  a description of a backward compatibility strategy that allows new
  servers to interact with the current HTML user agents.

  The proposal is independent of which version of HTML it becomes a
  part; in particular, no proposal is being made of whether this might
  be part of HTML 2.0, 2.1, or 3.0.

2.  HTML forms with file submission

  The current draft HTML specification <URL:http://www.hal.com/
  %7Fconnolly/html-spec/spyglass-19941014/html-19941014.txt.Z> defines
  eight possible values for the attribute TYPE of an INPUT element:
  CHECKBOX, HIDDEN, IMAGE, PASSWORD, RADIO, RESET, SUBMIT, TEXT.

  In addition, it defines the default ENCTYPE attribute of the FORM
  element using the POST METHOD to have the default value
  "application/x-www-form-urlencoded".

  This proposal makes three changes:
  1) add a FILE option for the TYPE attribute of INPUT
  2) Allow an ACCEPT attribute for INPUT tag, which is a list of
     media types or type patterns allowed for the input
  3) allow the ENCTYPE of a FORM to be "multipart/form-data".

  These changes might be considered independently, but are all
  necessary for reasonable file upload.

  The author of an HTML form who wants to request one or more files
  from a user would write (for example):

    <FORM ENCTYPE="multipart/form-data" ACTION="_URL_" METHOD=POST>

    File to process: <INPUT NAME="userfile1" TYPE="file">

    <INPUT TYPE="submit" VALUE="Send File">

    </FORM>

  The change to the HTML DTD is to add one item to the entity
  "InputType". In addition, it is proposed that the INPUT tag have an
  ACCEPT attribute, which is a list of comma-separated media types.


  ... (other elements) ...

  <!ENTITY % InputType "(TEXT | PASSWORD | CHECKBOX |
                         RADIO | SUBMIT | RESET |
                         IMAGE | HIDDEN | FILE )">
  <!ELEMENT INPUT - 0 EMPTY>
  <!ATTLIST INPUT
          TYPE %InputType TEXT
          NAME CDATA #IMPLIED  -- required for all but submit and reset
          VALUE CDATA #IMPLIED
          SRC %URI #IMPLIED  -- for image inputs --
          CHECKED (CHECKED) #IMPLIED
          SIZE CDATA #IMPLIED  --like NUMBERS,
                                  but delimited with comma, not space
          MAXLENGTH NUMBER #IMPLIED
          ALIGN (top|middle|bottom) #IMPLIED
          ACCEPT CDATA #IMPLIED --list of content types 
          >

  ... (other elements) ...

3.  Suggested implementation

  While user agents that interpret HTML have wide leeway to choose the
  most appropriate mechanism for their context, this section suggests
  how one class of user agent, WWW browsers, might implement file
  upload.

  When a INPUT tag of type FILE is encountered, the browser might show
  a display of (previously selected) file names, and a "Browse" button
  or selection method. Selecting the "Browse" button would cause the
  browser to enter into a file selection mode appropriate for the
  platform. Window-based browsers might pop up a file selection
  window, for example. In such a file selection dialog, the user would
  have the option of replacing a current selection, adding a new file
  selection, etc. Browser implementors might choose let the list of
  file names be manually edited.

  If an ACCEPT attribute is present, the browser might constrain the
  file patterns prompted for to match those with the corresponding
  appropriate file extensions for the platform.

  When the user completes the form, and selects the SUBMIT element,
  the browser should send the form data and the content of the
  selected files.  The encoding type application/x-www-form-urlencoded
  is inefficient for sending large quantities of binary data.  Thus, a
  new media type, multipart/form-data, is proposed as a way of
  efficiently sending the values associated with a filled-out form
  from client to server.

  The media-type multipart/form-data follows the rules of all
  multipart MIME data streams as outlined in RFC 1521--a boundary is
  selected that does not occur in any of the data.  Each field of the
  form is sent, in the order in which it occurs in the form, as a part
  of the multipart stream.  Each part identifies the INPUT name within
  the original HTML form using a "Name:" header.  Each part has an
  optional Content-Type (which defaults to text/plain). File inputs
  should be identified as either application/octet-stream or the
  appropriate media type, if known. If multiple files were selected,
  they should be transferred together using the multipart/mixed
  format.

  The "content-transfer-encoding" header should be supplied for all
  fields whose values do not conform to the default 7BIT encoding.
  (All characters 7-bit US-ASCII data with lines no longer than 1000
  characters.) Otherwise, file data and longer field values may be
  transferred using a content-transfer-encoding appropriate to the
  protocol of the ACTION in the form. For HTTP applications,
  content-transfer-encoding of "binary" may be use.  If the ACTION is
  a "mailto:" URL, then the user agent may encode the data
  appropriately to the mail transport mechanism.  [See section 5 of
  RFC 1521 for more details.]

  File inputs may optionally identify the file name using the
  "content-disposition" header. This is not required, but is as a
  convenience for those cases where, for example, the uploaded files
  might contain references to each other, e.g., a TeX file and its
  .sty auxiliary style description.

  Browers may optionally include a Content-Length header both in the
  overall reply and in individual components. The content-length is
  *not* intended as a replacement for the multipart boundary as a way
  of detecting the end of an individual component. It is *only*
  supplied as a way forwarning the server of the amount of data
  coming.
 
  On the server end, the ACTION might point to a HTTP URL that
  implements the forms action via CGI. In such a case, the CGI program
  would note that the content-type is multipart/form-data, parse the
  various fields (checking for validity, writing the file data to local
  files for subsequent processing, etc.).

4.  Backward compatibility issues

  While not necessary for successful adoption of an enhancement to the
  current WWW form mechanism, it is useful to also plan for a
  migration strategy: users with older browsers can still participate
  in file upload dialogs, using a helper application. Most current web
  browers, when given <INPUT TYPE=FILE>, will treat it as <INPUT
  TYPE=TEXT> and give the user a text box. The user can type in a file
  name into this text box. In addition, current browsers seem to
  ignore the ENCTYPE parameter in the <FORM> element, and always
  transmit the data as application/x-www-form-urlencoded.

  Thus, the server CGI might be written in a way that would note that
  the form data returned had content-type
  application/x-www-form-urlencoded instead of
  multipart/form-data, and know that the user was using a browser
  that didn't implement file upload.

  In this case, rather than replying with a "text/html" response, the
  CGI on the server could instead send back a data stream that a helper
  application might process instead; this would be a data stream of
  type "application/x-please-send-files", which contains: 

  * The (fully qualified) URL to which the actual form data should
    be posted (terminated with CRLF)
  * The list of field names that were supposed to be file contents
    (space separated, terminated with CRLF)
  * The entire original application/x-www-form-urlencoded form data
    as originally sent from client to server.
    
  In this case, the browser needs to be configured to process
  application/x-please-send-files to launch a helper application.

  The helper would read the form data, note which fields contained
  'local file names' that needed to be replaced with their data
  content, might itself prompt the user for changing or adding to the
  list of files available, and then repackage the data & file contents
  in multipart/form-data for retransmission back to the server.

  The helper would generate the kind of data that a 'new' browser should
  actually have sent in the first place, with the intention that the URL
  to which it is sent corresponds to the original ACTION URL. The point
  of this is that the server can use the *same* CGI to implement the
  mechanism for dealing with both old and new browsers.

  The helper need not display the form data, but *should* ensure that
  the user actually be prompted about the suitability of sending the
  files requested (this is to avoid a security problem with malicious
  servers that ask for files that weren't actually promised by the
  user.) It would be useful if the status of the transfer of the files
  involved could be displayed.

5.  Other considerations

5.1 Compression, encryption

  This scheme doesn't address the possible compression of files.
  After some consideration, it seemed that the optimization issues of
  file compression were too complex to try to automatically have
  browsers decide that files should be compressed.  Many link-layer
  transport mechanisms (e.g., high-speed modems) perform data
  compression over the link, and optimizing for compression at this
  layer might not be appropriate. It might be possible for browsers to
  optionally produce a content-transfer-encoding of x-compress for
  file data, and for servers to decompress the data before processing,
  if desired; this was left out of the proposal, however.

  Similarly, the proposal does not contain a mechanism for encryption
  of the data; this should be handled by whatever other mechanisms are
  in place for secure transmission of data, whether via secure HTTP or
  mail.

5.2 Deferred file transmission

  In some situations, it might be advisable to have the server
  validate various elements of the form data (user name, account,
  etc.)  before actually preparing to receive the data.  However,
  after some consideration, it seemed best to require that servers
  that wish to do this should implement this as a series of forms,
  where some of the data elements that were previously validated might
  be sent back to the client as 'hidden' fields, or by arranging the
  form so that the elements that need validation occur first.  This
  puts the onus of maintaining the state of a transaction only on
  those servers that wish to build a complex application, while
  allowing those cases that have simple input needs to be built
  simply. Clients are encouraged to supply content-length for overall
  file input so that a busy server could detect if the proposed file
  data is too large to be processed reasonably and just return an
  error code and close the connection without waiting to process all
  of the incoming data.

  If the INPUT tag includes the attribute MAXLENGTH, the user agent
  should consider its value to represent the maximum Content-Length
  (in bytes) which the server will accept for transferred files.  In
  this way, servers can hint to the client how much space they have
  available for a file upload, before that upload takes place.  It is
  important to note, however, that this is only a hint, and the actual
  requirements of the server may change between form creation and file
  submission.

5.3 Other choices for return transmission of binary data

  Various people have suggested using new mime top-level type
  "aggregate", e.g., aggregate/mixed or a content-transfer-encoding of
  "packet" to express indeterminate-length binary data, rather than
  relying on the multipart-style boundaries.  While we are not opposed
  to doing so, this would require additional design and
  standardization work to get acceptance of "aggregate".  On the other
  hand, the 'multipart' mechanisms are well established, simple to
  implement on both the sending client and receiving server, and as
  efficient as other methods of dealing with multiple combinations of
  binary data.

5.4 Not overloading <INPUT>:

  Various people have wondered about the advisability of overloading
  'INPUT' for this function, rather than merely providing a different
  type of FORM element.  Among other considerations, the migration
  strategy which is allowed when using <INPUT> is important.  In
  addition, the <INPUT> field *is* already overloaded to contain most
  kinds of data input; rather than creating multiple kinds of <INPUT>
  tags, it seems most reasonable to enhance <INPUT>.  The 'type' of
  INPUT is not the content-type of what is returned, but rather the
  'widget-type'; i.e., it identifies the interaction style with the
  user.  The description here is carefully written to allow <INPUT
  TYPE=FILE> to work for text browsers or audio-markup.

5.5 Default content-type of field data

  Many input fields in HTML are to be typed in. There has been some
  ambiguity as to how form data should be transmitted back to servers.
  Making the content-type of <INPUT> fields be text/plain clearly
  disambiguates that the client should properly encode the data before
  sending it back to the server with CRLFs.

5.6 Interpretation of other attributes

  The VALUE attribute might be used with <INPUT TYPE=file> tags for
  a default file name. This use is probably platform dependent,
  however, and probably should be avoided.

  The SIZE attribute should probably not be used with <INPUT
  TYPE=file> tags.  For text input, it indicates the length in
  characters for the text area for the prompt. 

5.7 Using Name instead of existing MIME headers for field names

  To avoid any confusion and overloading of existing mime headers,
  this proposal has "Name" (instead of content-id or
  content-description) for the name of the original field in the form
  in multipart/form-data. 

  On the other hand, the original file names might appear within a
  multipart/mixed stream, and thus the proposal avoids defining a new
  header field for transmitting the original file name by using
  the "content-disposition" field to send the file name, if
  appropriate. (Content-disposition is apparently the proposed way of
  encoding file names in MIME messages, as well.)

5.8 Allow form ACTION to be "mailto:"

  Independent of this proposal, it would be very useful for HTML
  interpreting user agents to allow a ACTION in a form to be a
  "mailto:" URL. This seems like a good idea, with or without this
  proposal. Similarly, the ACTION for a HTML form which is received
  via mail should probably default to the "reply-to:" of the message.
  These two proposals would allow HTML forms to be served via HTTP
  servers but sent back via mail, or, alternatively, allow HTML forms
  to be sent by mail, filled out by HTML-aware mail recipients, and
  the results mailed back.
  
5.9 Remote files with third-party transfer

  In some scenarios, the user operating the client software might want
  to specify a URL for remote data rather than a local file. In this
  case, is there a way to allow the browser to send to the client a
  pointer to the external data rather than the entire contents? This
  capability could be implemented, for example, by having the client
  send to the server data of type "message/external-body" with
  "access-type" set to, say, "uri", and the URL of the remote data in
  the body of the message.

5.10 File transfer with ENCTYPE=x-www-form-urlencoded

  If a form contains <INPUT TYPE=file> elements but does not contain
  an ENCTYPE in the enclosing <FORM>, the behavior is not specified.
  It is probably inappropriate to attempt to URN-encode large
  quantities of data to servers that don't expect it.

5.11 CRLF used as line separator

  As with all MIME transmissions, CRLF is used as the separator for
  lines in a POST of the data in multipart/www-form-data.

6. Examples

   Suppose the server supplies the following HTML:

     <FORM ACTION="http://server.dom/cgi/handle"
           ENCTYPE="multipart/form-data"
           METHOD=POST>
     What is your name? <INPUT TYPE=TEXT NAME=submitter>
     What files are you sending? <INPUT TYPE=FILE NAME=pics>
     </FORM>

   and the user types "Joe Blow" in the name field, and selects
   a text file "file1.txt" and also an image file "file2.gif" for
   the answer to 'What files are you sending?'.

   The client would send back the following data:

   Content-type: multipart/form-data, boundary=AaB03x
   --AaB03x
   Name: field1

   Joe Blow
   --AaB03x
   Name: pics
   Content-type: multipart/mixed, boundary=BbC04y

   --BbC04y
   Content-Description: file1.txt
   Content-Type: text/plain
   Content-Transfer-Encoding: binary

   ... contents of file1.txt ...
   --BbC04y
   Content-Description: file2.gif
   Content-type: image/gif
   Content-Transfer-Encoding: binary
   
   ...contents of file2.gif...
   --BbC04y--
   --AaB03x--


7.  Conclusion

  The suggested implementation gives the client a lot of flexibility in
  the number and types of files it can send to the server, it gives the
  server control of the decision to accept the files, and it gives
  servers a chance to interact with browsers which do not support INPUT
  TYPE "file".

  The change to the HTML DTD is very simple, but very powerful.  It
  enables a much greater variety of services to be implemented via the
  World-Wide Web than is currently possible due to the lack of a file
  submission facility.  This would be an extremely valuable addition to
  the capabilities of the World-Wide Web.


A.  Authors' Addresses

  Larry Masinter                          masinter@parc.xerox.com
  Xerox Palo Alto Research Center         Voice:  (415) 812-4365
  3333 Coyote Hill Road                   Fax:    (415) 812-4333
  Palo Alto, CA 94304

  Ernesto Nebel                           nebel@xsoft.sd.xerox.com
  XSoft, Xerox Corporation                Voice:  (619) 676-7817
  10875 Rancho Bernardo Road, Suite 200   Fax:    (619) 676-7865
  San Diego, CA 92127-2116

1
2	INTERNET-DRAFT E. Nebel
3	Form-based File Upload in HTML L. Masinter
4	draft-ietf-html-fileupload-01.txt Xerox Corporation
5	Expires May 30, 1995 November 30, 1994
6
7
8	Form-based File Upload in HTML
9
10	Status of this Memo
11
12	This document is an Internet-Draft. Internet-Drafts are working
13	documents of the Internet Engineering Task Force (IETF), its areas,
14	and its working groups. Note that other groups may also distribute
15	working documents as Internet-Drafts.
16
17	Internet-Drafts are draft documents valid for a maximum of six months
18	and may be updated, replaced, or obsoleted by other documents at any
19	time. It is inappropriate to use Internet-Drafts as reference
20	material or to cite them other than as ``work in progress.''
21
22	To learn the current status of any Internet-Draft, please check the
23	``1id-abstracts.txt'' listing contained in the Internet-Drafts Shadow
24	Directories on ftp.is.co.za (Africa), nic.nordu.net (Europe),
25	munnari.oz.au (Pacific Rim), ds.internic.net (US East Coast), or
26	ftp.isi.edu (US West Coast).
27
28
29	1. Abstract
30
31	Currently, HTML forms allow the producer of the form to request
32	information from the user reading the form. These forms have proven
33	useful in a wide variety of applications in which input from the
34	user is necessary. However, this capability is limited because HTML
35	forms don't provide a way to ask the user to submit files of data.
36	Service providers who need to get files from the user have had to
37	implement custom user applications. (Examples of these custom
38	browsers have appeared on the www-talk mailing list.) Since
39	file-upload is a feature that will benefit many applications, this
40	draft proposes an extension to HTML to allow information providers
41	to express file upload requests uniformly, and a MIME compatible
42	representation for file upload responses. This draft also includes
43	a description of a backward compatibility strategy that allows new
44	servers to interact with the current HTML user agents.
45
46	The proposal is independent of which version of HTML it becomes a
47	part; in particular, no proposal is being made of whether this might
48	be part of HTML 2.0, 2.1, or 3.0.
49
50	2. HTML forms with file submission
51
52	The current draft HTML specification <URL:http://www.hal.com/
53	%7Fconnolly/html-spec/spyglass-19941014/html-19941014.txt.Z> defines
54	eight possible values for the attribute TYPE of an INPUT element:
55	CHECKBOX, HIDDEN, IMAGE, PASSWORD, RADIO, RESET, SUBMIT, TEXT.
56
57	In addition, it defines the default ENCTYPE attribute of the FORM
58	element using the POST METHOD to have the default value
59	"application/x-www-form-urlencoded".
60
61	This proposal makes three changes:
62	1) add a FILE option for the TYPE attribute of INPUT
63	2) Allow an ACCEPT attribute for INPUT tag, which is a list of
64	media types or type patterns allowed for the input
65	3) allow the ENCTYPE of a FORM to be "multipart/form-data".
66
67	These changes might be considered independently, but are all
68	necessary for reasonable file upload.
69
70	The author of an HTML form who wants to request one or more files
71	from a user would write (for example):
72
73	<FORM ENCTYPE="multipart/form-data" ACTION="_URL_" METHOD=POST>
74
75	File to process: <INPUT NAME="userfile1" TYPE="file">
76
77	<INPUT TYPE="submit" VALUE="Send File">
78
79	</FORM>
80
81	The change to the HTML DTD is to add one item to the entity
82	"InputType". In addition, it is proposed that the INPUT tag have an
83	ACCEPT attribute, which is a list of comma-separated media types.
84
85
86	... (other elements) ...
87
88	<!ENTITY % InputType "(TEXT \| PASSWORD \| CHECKBOX \|
89	RADIO \| SUBMIT \| RESET \|
90	IMAGE \| HIDDEN \| FILE )">
91	<!ELEMENT INPUT - 0 EMPTY>
92	<!ATTLIST INPUT
93	TYPE %InputType TEXT
94	NAME CDATA #IMPLIED -- required for all but submit and reset
95	VALUE CDATA #IMPLIED
96	SRC %URI #IMPLIED -- for image inputs --
97	CHECKED (CHECKED) #IMPLIED
98	SIZE CDATA #IMPLIED --like NUMBERS,
99	but delimited with comma, not space
100	MAXLENGTH NUMBER #IMPLIED
101	ALIGN (top\|middle\|bottom) #IMPLIED
102	ACCEPT CDATA #IMPLIED --list of content types
103	>
104
105	... (other elements) ...
106
107	3. Suggested implementation
108
109	While user agents that interpret HTML have wide leeway to choose the
110	most appropriate mechanism for their context, this section suggests
111	how one class of user agent, WWW browsers, might implement file
112	upload.
113
114	When a INPUT tag of type FILE is encountered, the browser might show
115	a display of (previously selected) file names, and a "Browse" button
116	or selection method. Selecting the "Browse" button would cause the
117	browser to enter into a file selection mode appropriate for the
118	platform. Window-based browsers might pop up a file selection
119	window, for example. In such a file selection dialog, the user would
120	have the option of replacing a current selection, adding a new file
121	selection, etc. Browser implementors might choose let the list of
122	file names be manually edited.
123
124	If an ACCEPT attribute is present, the browser might constrain the
125	file patterns prompted for to match those with the corresponding
126	appropriate file extensions for the platform.
127
128	When the user completes the form, and selects the SUBMIT element,
129	the browser should send the form data and the content of the
130	selected files. The encoding type application/x-www-form-urlencoded
131	is inefficient for sending large quantities of binary data. Thus, a
132	new media type, multipart/form-data, is proposed as a way of
133	efficiently sending the values associated with a filled-out form
134	from client to server.
135
136	The media-type multipart/form-data follows the rules of all
137	multipart MIME data streams as outlined in RFC 1521--a boundary is
138	selected that does not occur in any of the data. Each field of the
139	form is sent, in the order in which it occurs in the form, as a part
140	of the multipart stream. Each part identifies the INPUT name within
141	the original HTML form using a "Name:" header. Each part has an
142	optional Content-Type (which defaults to text/plain). File inputs
143	should be identified as either application/octet-stream or the
144	appropriate media type, if known. If multiple files were selected,
145	they should be transferred together using the multipart/mixed
146	format.
147
148	The "content-transfer-encoding" header should be supplied for all
149	fields whose values do not conform to the default 7BIT encoding.
150	(All characters 7-bit US-ASCII data with lines no longer than 1000
151	characters.) Otherwise, file data and longer field values may be
152	transferred using a content-transfer-encoding appropriate to the
153	protocol of the ACTION in the form. For HTTP applications,
154	content-transfer-encoding of "binary" may be use. If the ACTION is
155	a "mailto:" URL, then the user agent may encode the data
156	appropriately to the mail transport mechanism. [See section 5 of
157	RFC 1521 for more details.]
158
159	File inputs may optionally identify the file name using the
160	"content-disposition" header. This is not required, but is as a
161	convenience for those cases where, for example, the uploaded files
162	might contain references to each other, e.g., a TeX file and its
163	.sty auxiliary style description.
164
165	Browers may optionally include a Content-Length header both in the
166	overall reply and in individual components. The content-length is
167	not intended as a replacement for the multipart boundary as a way
168	of detecting the end of an individual component. It is only
169	supplied as a way forwarning the server of the amount of data
170	coming.
171
172	On the server end, the ACTION might point to a HTTP URL that
173	implements the forms action via CGI. In such a case, the CGI program
174	would note that the content-type is multipart/form-data, parse the
175	various fields (checking for validity, writing the file data to local
176	files for subsequent processing, etc.).
177
178	4. Backward compatibility issues
179
180	While not necessary for successful adoption of an enhancement to the
181	current WWW form mechanism, it is useful to also plan for a
182	migration strategy: users with older browsers can still participate
183	in file upload dialogs, using a helper application. Most current web
184	browers, when given <INPUT TYPE=FILE>, will treat it as <INPUT
185	TYPE=TEXT> and give the user a text box. The user can type in a file
186	name into this text box. In addition, current browsers seem to
187	ignore the ENCTYPE parameter in the <FORM> element, and always
188	transmit the data as application/x-www-form-urlencoded.
189
190	Thus, the server CGI might be written in a way that would note that
191	the form data returned had content-type
192	application/x-www-form-urlencoded instead of
193	multipart/form-data, and know that the user was using a browser
194	that didn't implement file upload.
195
196	In this case, rather than replying with a "text/html" response, the
197	CGI on the server could instead send back a data stream that a helper
198	application might process instead; this would be a data stream of
199	type "application/x-please-send-files", which contains:
200
201	* The (fully qualified) URL to which the actual form data should
202	be posted (terminated with CRLF)
203	* The list of field names that were supposed to be file contents
204	(space separated, terminated with CRLF)
205	* The entire original application/x-www-form-urlencoded form data
206	as originally sent from client to server.
207
208	In this case, the browser needs to be configured to process
209	application/x-please-send-files to launch a helper application.
210
211	The helper would read the form data, note which fields contained
212	'local file names' that needed to be replaced with their data
213	content, might itself prompt the user for changing or adding to the
214	list of files available, and then repackage the data & file contents
215	in multipart/form-data for retransmission back to the server.
216
217	The helper would generate the kind of data that a 'new' browser should
218	actually have sent in the first place, with the intention that the URL
219	to which it is sent corresponds to the original ACTION URL. The point
220	of this is that the server can use the same CGI to implement the
221	mechanism for dealing with both old and new browsers.
222
223	The helper need not display the form data, but should ensure that
224	the user actually be prompted about the suitability of sending the
225	files requested (this is to avoid a security problem with malicious
226	servers that ask for files that weren't actually promised by the
227	user.) It would be useful if the status of the transfer of the files
228	involved could be displayed.
229
230	5. Other considerations
231
232	5.1 Compression, encryption
233
234	This scheme doesn't address the possible compression of files.
235	After some consideration, it seemed that the optimization issues of
236	file compression were too complex to try to automatically have
237	browsers decide that files should be compressed. Many link-layer
238	transport mechanisms (e.g., high-speed modems) perform data
239	compression over the link, and optimizing for compression at this
240	layer might not be appropriate. It might be possible for browsers to
241	optionally produce a content-transfer-encoding of x-compress for
242	file data, and for servers to decompress the data before processing,
243	if desired; this was left out of the proposal, however.
244
245	Similarly, the proposal does not contain a mechanism for encryption
246	of the data; this should be handled by whatever other mechanisms are
247	in place for secure transmission of data, whether via secure HTTP or
248	mail.
249
250	5.2 Deferred file transmission
251
252	In some situations, it might be advisable to have the server
253	validate various elements of the form data (user name, account,
254	etc.) before actually preparing to receive the data. However,
255	after some consideration, it seemed best to require that servers
256	that wish to do this should implement this as a series of forms,
257	where some of the data elements that were previously validated might
258	be sent back to the client as 'hidden' fields, or by arranging the
259	form so that the elements that need validation occur first. This
260	puts the onus of maintaining the state of a transaction only on
261	those servers that wish to build a complex application, while
262	allowing those cases that have simple input needs to be built
263	simply. Clients are encouraged to supply content-length for overall
264	file input so that a busy server could detect if the proposed file
265	data is too large to be processed reasonably and just return an
266	error code and close the connection without waiting to process all
267	of the incoming data.
268
269	If the INPUT tag includes the attribute MAXLENGTH, the user agent
270	should consider its value to represent the maximum Content-Length
271	(in bytes) which the server will accept for transferred files. In
272	this way, servers can hint to the client how much space they have
273	available for a file upload, before that upload takes place. It is
274	important to note, however, that this is only a hint, and the actual
275	requirements of the server may change between form creation and file
276	submission.
277
278	5.3 Other choices for return transmission of binary data
279
280	Various people have suggested using new mime top-level type
281	"aggregate", e.g., aggregate/mixed or a content-transfer-encoding of
282	"packet" to express indeterminate-length binary data, rather than
283	relying on the multipart-style boundaries. While we are not opposed
284	to doing so, this would require additional design and
285	standardization work to get acceptance of "aggregate". On the other
286	hand, the 'multipart' mechanisms are well established, simple to
287	implement on both the sending client and receiving server, and as
288	efficient as other methods of dealing with multiple combinations of
289	binary data.
290
291	5.4 Not overloading <INPUT>:
292
293	Various people have wondered about the advisability of overloading
294	'INPUT' for this function, rather than merely providing a different
295	type of FORM element. Among other considerations, the migration
296	strategy which is allowed when using <INPUT> is important. In
297	addition, the <INPUT> field is already overloaded to contain most
298	kinds of data input; rather than creating multiple kinds of <INPUT>
299	tags, it seems most reasonable to enhance <INPUT>. The 'type' of
300	INPUT is not the content-type of what is returned, but rather the
301	'widget-type'; i.e., it identifies the interaction style with the
302	user. The description here is carefully written to allow <INPUT
303	TYPE=FILE> to work for text browsers or audio-markup.
304
305	5.5 Default content-type of field data
306
307	Many input fields in HTML are to be typed in. There has been some
308	ambiguity as to how form data should be transmitted back to servers.
309	Making the content-type of <INPUT> fields be text/plain clearly
310	disambiguates that the client should properly encode the data before
311	sending it back to the server with CRLFs.
312
313	5.6 Interpretation of other attributes
314
315	The VALUE attribute might be used with <INPUT TYPE=file> tags for
316	a default file name. This use is probably platform dependent,
317	however, and probably should be avoided.
318
319	The SIZE attribute should probably not be used with <INPUT
320	TYPE=file> tags. For text input, it indicates the length in
321	characters for the text area for the prompt.
322
323	5.7 Using Name instead of existing MIME headers for field names
324
325	To avoid any confusion and overloading of existing mime headers,
326	this proposal has "Name" (instead of content-id or
327	content-description) for the name of the original field in the form
328	in multipart/form-data.
329
330	On the other hand, the original file names might appear within a
331	multipart/mixed stream, and thus the proposal avoids defining a new
332	header field for transmitting the original file name by using
333	the "content-disposition" field to send the file name, if
334	appropriate. (Content-disposition is apparently the proposed way of
335	encoding file names in MIME messages, as well.)
336
337	5.8 Allow form ACTION to be "mailto:"
338
339	Independent of this proposal, it would be very useful for HTML
340	interpreting user agents to allow a ACTION in a form to be a
341	"mailto:" URL. This seems like a good idea, with or without this
342	proposal. Similarly, the ACTION for a HTML form which is received
343	via mail should probably default to the "reply-to:" of the message.
344	These two proposals would allow HTML forms to be served via HTTP
345	servers but sent back via mail, or, alternatively, allow HTML forms
346	to be sent by mail, filled out by HTML-aware mail recipients, and
347	the results mailed back.
348
349	5.9 Remote files with third-party transfer
350
351	In some scenarios, the user operating the client software might want
352	to specify a URL for remote data rather than a local file. In this
353	case, is there a way to allow the browser to send to the client a
354	pointer to the external data rather than the entire contents? This
355	capability could be implemented, for example, by having the client
356	send to the server data of type "message/external-body" with
357	"access-type" set to, say, "uri", and the URL of the remote data in
358	the body of the message.
359
360	5.10 File transfer with ENCTYPE=x-www-form-urlencoded
361
362	If a form contains <INPUT TYPE=file> elements but does not contain
363	an ENCTYPE in the enclosing <FORM>, the behavior is not specified.
364	It is probably inappropriate to attempt to URN-encode large
365	quantities of data to servers that don't expect it.
366
367	5.11 CRLF used as line separator
368
369	As with all MIME transmissions, CRLF is used as the separator for
370	lines in a POST of the data in multipart/www-form-data.
371
372	6. Examples
373
374	Suppose the server supplies the following HTML:
375
376	<FORM ACTION="http://server.dom/cgi/handle"
377	ENCTYPE="multipart/form-data"
378	METHOD=POST>
379	What is your name? <INPUT TYPE=TEXT NAME=submitter>
380	What files are you sending? <INPUT TYPE=FILE NAME=pics>
381	</FORM>
382
383	and the user types "Joe Blow" in the name field, and selects
384	a text file "file1.txt" and also an image file "file2.gif" for
385	the answer to 'What files are you sending?'.
386
387	The client would send back the following data:
388
389	Content-type: multipart/form-data, boundary=AaB03x
390	--AaB03x
391	Name: field1
392
393	Joe Blow
394	--AaB03x
395	Name: pics
396	Content-type: multipart/mixed, boundary=BbC04y
397
398	--BbC04y
399	Content-Description: file1.txt
400	Content-Type: text/plain
401	Content-Transfer-Encoding: binary
402
403	... contents of file1.txt ...
404	--BbC04y
405	Content-Description: file2.gif
406	Content-type: image/gif
407	Content-Transfer-Encoding: binary
408
409	...contents of file2.gif...
410	--BbC04y--
411	--AaB03x--
412
413
414	7. Conclusion
415
416	The suggested implementation gives the client a lot of flexibility in
417	the number and types of files it can send to the server, it gives the
418	server control of the decision to accept the files, and it gives
419	servers a chance to interact with browsers which do not support INPUT
420	TYPE "file".
421
422	The change to the HTML DTD is very simple, but very powerful. It
423	enables a much greater variety of services to be implemented via the
424	World-Wide Web than is currently possible due to the lack of a file
425	submission facility. This would be an extremely valuable addition to
426	the capabilities of the World-Wide Web.
427
428
429	A. Authors' Addresses
430
431	Larry Masinter masinter@parc.xerox.com
432	Xerox Palo Alto Research Center Voice: (415) 812-4365
433	3333 Coyote Hill Road Fax: (415) 812-4333
434	Palo Alto, CA 94304
435
436	Ernesto Nebel nebel@xsoft.sd.xerox.com
437	XSoft, Xerox Corporation Voice: (619) 676-7817
438	10875 Rancho Bernardo Road, Suite 200 Fax: (619) 676-7865
439	San Diego, CA 92127-2116
440