6 ** Better implementation for multipart/alternative
8 ** dynamic configuration for 'mime-preview-condition
10 ** Don't use filter-model
12 tomo (major developer of SEMI) and akr (major developer of
13 FLIM-FLAM) discussed about Emacs 20.3 problem related with SEMI and
14 FLIM. They found essential problem with Emacs 20.3 are:
16 - Emacs 20.3 separates string-type to unibyte-string and
17 multibyte-string. Emacs 20.3 has enough APIs to treat them.
19 - Buffer has mode about interpretation of contents. Each mode is
20 designed to save semantics as characters. Namely buffer contains
21 unibyte-characters or multibyte-characters. One buffer can not
22 contain both representations.
24 - {decode|encode}-coding-{region|string} run in byte world. So it
25 is not harmonized with other API.
27 - It seems easy to write code in one mode or one world
28 (unibyte-string or multibyte-string). However it seems not easy
29 to write inter-mode program, such as SEMI.
31 - Byte <-> byte conversion, such as base64, quoted-printable, must
32 be run only with unibyte-mode.
34 - Byte <-> character conversion, such as
35 {decode|encode}-coding-region, should not be defined in single
36 buffer. Instead of them, decoder should read from unibyte buffer
37 and output to multibyte buffer. Similarly, encoder should read
38 from multibyte buffer and output to unibyte buffer.
39 `insert-buffer-substring' like API may be suitable. Anyway Emacs
40 introduces multiple representations, so it should be redesigned
41 based on multiple representation world model.
43 Anyway FLIM should introduce new APIs based on inter-representation
44 world model. Conventional APIs should be implemented based on new
47 SEMI should abolish filter model and introduce new methods to
48 display inline data. These methods should use new FLIM APIs based on
49 inter-representation world model.
54 ** WYSIWYG editing support
56 ** Use MIME-Preview like tag and display
58 ** Redesign to use two buffers for one message
60 MIME-View is based on "Multiple Representation Space (layer) Model".
61 In this model, network representation and its presentation are
62 distinguished. Thus MIME-View uses two buffers for one message,
63 'mime-raw-buffer (for network representation) and
64 'mime-preview-buffer. MIME-View manages them based on information of
65 entities. According to experience of MIME-View, this model is good to
66 treat complex structured data, such as MIME.
68 MIME-Edit was designed to use one buffer for one message. So it is
69 hard to edit like WYSIWYG style. Format of tag is limited by
70 translation. Content of forwarded message is unreadable. It is
71 better to introduce "Multiple Representation Space Model" to resolve
74 ** Check available MIME-charset
76 MIME-charset
\e$B0J30$,@8@.$5$l$k>l9g$N=hM}$r;XDj$G$-$k$h$&$K$9$k!#
\e(B
80 (a) translate problematic characters to similar representation
81 (b) display warning message
82 (e.g. "`x-ctext' is generated. Do you send it? (yes/no)")
85 ** Don't use buffer-local variables
87 Don't use buffer-local variables to control behavior about
88 translating to network representation, such as 'mime-transfer-level,
89 'mime-transfer-level-string,
90 'mime-edit-charset-default-encoding-alist, 'mime-edit-pgp-processing.
91 Because they have problem with Semi-gnus.