The Origin of Blockquote

This is a partial log of the IRC channel #swig on
from 2017-08-28, with timestamps in UTC, where Dan Connolly and I
discuss the origin of the blockquote element in HTML circa December
1992. This channel is normally archived and accessible to the public,
but the logger has not been there since February 2017.

The opening line of the log refers to a comment by DanC that he had
found his 1991 mail archive from when he was at Convex, in the wider
context of trying to rediscover more about the origins of HTML.

18:02:31 <sbp> DanC: what does your 1991 Convex mail contain? can it
be made public?
18:03:44 <DanC_> I copied www-talk on stuff that, at the time, I
though should be public, and I haven't seen anything where I'd change
my mind
18:04:56 <DanC_> the W3C systems guys sent me all the ,v files.
18:06:50 <DanC_> so I have all the versions; unfortunately, when I
imported my Convex work into the W3C CVS repository in 1994, the
original dates weren't preserved.
18:09:39 <sbp> DanC_: ah, shame. so there's no extra information there
as to why you invented blockquote? last time we spoke about it, in
2013, you thought you might have been inspired by LaTeX. but we could
only find \begin{quote}, which didn't look to either of us as though
it was a likely precursor
18:15:23 <DanC_> blockquote was added in "The DTD as I originally
released it." but the date of that rev is goofy
18:16:04 <DanC_> around 1st appearance of BLOCKQUOTE in www-talk, we
were talking about texinfo; does it have blockquote?
18:17:05 <sbp> checking
18:18:05 <DanC_>
18:18:07 <DanC_> not exactly
18:18:08 <DanC_> • @quotation:         Writing a quotation.
18:18:08 <DanC_> • @indentedblock:         Block of text indented on left.
18:21:57 <DanC_> hard to imagine I made it up. have you looked at
linemode browser code?
18:22:06 <DanC_> or Mosaic?
18:22:47 <sbp> I got
ostensibly dated to 1994-07-22 and grepped for -i block and -i quote,
but nothing obviously relevant came up
18:22:49 <DanC_> ooh... nice...
18:23:21 <sbp> DanC_: I did some analysis of early HTML, yeah. I
actually compiled a corpus of early HTML pages and then did a plot
based on element popularities
18:23:56 <DanC_> pointer?
18:24:02 <sbp>
18:24:21 <sbp> (quite old work)
18:24:35 <sbp> but I think blockquote postdated the range I was
looking at, because it doesn't appear
18:25:03 <sbp> the earliest references to it that I found remain those
documented in my 2009 email to you,
18:25:26 <sbp> namely, and
v1.3 of html.dtd
18:26:30 <DanC_> wild... looks like*checkout*/test/html.dtd  is what I've
been hunting for
18:26:32 <sbp> having said that, you do introduce the idea of
BLOCKQUOTE with "I'm trying to keep up with all sorts of HTML ideas"
18:26:42 <sbp> which does suggest that you were inspired by somebody else
18:27:08 <sbp> but to date I have not found any antedating of blockquote in HTML
18:30:47 <DanC_> 2010 discussion of csquotes and spacing with
18:33:14 <sbp> right, but HTML has had such wide influence that
anything post-early-1990s is tainted by its popularity. it's most
likely that any use of the word "blockquote" after then is directly
inspired by HTML
18:33:43 <sbp> (which is why e.g. I want back and looked for
texinfo-3.1.tar.gz rather than reading the present manual)
18:34:32 <sbp> that little check-in that you did to html.dtd on 7th
Jan 1993 has changed the English language forever! :-)
18:35:27 <DanC_> when did Mosaic pick it up?
18:35:35 <DanC_> I wonder about MidasWWW...
18:35:38 <sbp> good question, investigating
18:36:34 <DanC_> MidasWWW Nov 16, 1992 did not have it
18:39:51 <DanC_> wow... there are just no other copies of MidasWWW to
be had... TimBL eventually got pretty good at caching his own copies
of things, but I wonder at what point
18:40:21 <sbp> the oldest release of Mosaic that I can find is
(0.6b), which contains WMOSAIC.EXE. when I use strings and grep on it,
I get various responses for blockquote
18:40:31 <sbp> e.g. "BLOCKQUOTE" and "BlockQuote Font"
18:40:52 <sbp> according to that
was released on 28 Sep 1993
18:41:04 <sbp> long after the first mention in www-talk
18:41:24 <sbp> indeed, 0.1a was developed in June 1993 so that's too late too
18:41:53 <DanC_> hmm... I wonder if the next browser added it
18:42:02 <DanC_> surely the linemode browser did...
18:42:28 DanC_ hopes goes all the way back...
18:44:01 <sbp> I've got NextStep 0.16 here dated 1 Feb 1993—that
should be helpful
18:44:22 <sbp> doesn't contain "blockquote"
18:44:23 <DanC_> Author: Tim Berners-Lee <>  1993-02-03 03:52:58
18:45:05 <sbp> ah, here we go:
18:45:06 <sbp> ./Implementation/DefaultStyles.c:    { &HTStyleAddress,
 "BlockQuote", "BLOCKQUOTE",
18:45:08 <DanC_> gr... what's the git equivalent of hg grep? git grep
doesn't show versions
18:45:12 <sbp> 11 May  1993 ./Implementation/DefaultStyles.c
18:45:47 <sbp> SO says "git grep <regexp> $(git rev-list --all)"
18:45:54 <sbp> intuitive, eh?
18:45:58 <DanC_> oh. of course. obviously.
18:48:44 <DanC_> so... what was going on around 7th Jan 1993? I was
still at Convex near DFW...
18:48:47 <sbp> this was from linemode v2.11
18:49:02 <sbp> I've tried looking for v2.10 with no success yet, but I
did find the change log here:
18:49:16 <sbp> and under v2.10 it says, simply, " Millions of things
which seem to have slipped past this list in versions 2.x to s.09,
inclduing the rule file, firing off X applications for graphics files,
etc etc. See the library change list."
18:50:19 <sbp> there's also
18:50:36 <sbp> that mentions, on 11 Dec 1992, "Added <PRE> tag as in
new HTML spec."
18:50:41 <sbp> but nothing about <BLOCKQUOTE>!
18:50:57 <DanC_> I was working on FrameMaker support for HTML I think...
18:51:06 <DanC_> make FrameMaker MML had BLOCKQUOTE?
18:51:11 <DanC_> s/make/maybe/
18:51:15 <DanC_>
18:52:05 <DanC_>
18:53:09 <sbp> links to which is
18:53:31 <sbp> but it's in
18:53:40 <sbp> i.e.
18:53:48 <DanC_> ew... evil: <- is also 404
18:54:00 <sbp> only mention of "block" or "quote" is "; quote is short
for IN, i.e. inch"
18:56:24 <sbp> ah,
18:56:35 <sbp> this is from 21 Jan 1993 and contains blockquote
18:56:57 <sbp> libHTML.a in the same directory is from 29 Jan 1993 and
also contains blockquote
19:00:30 <DanC_> ah... while looking around my Convex mail, I find my
Mosaic patch for HTML parsing...
19:00:39 <DanC_> not following up on that is one of my great regrets.
19:02:55 <DanC_> yes... I was leaving Convex in Jan '93. Turbulent time...
19:03:14 <DanC_> had just gotten engaged
19:04:47 <ww> why did we stop using uppercase tags?
19:07:10 <sbp> ah,
mentions blockquote
19:07:27 <sbp> but only in the context of the DTD
19:07:38 <sbp> and since this is dated 11 Jan...
19:07:59 <sbp> it's very close to the 7th Jan check in for blockquote
19:09:11 <sbp> timbl mentions BLOCKQUOTE in his reply to you, but he
doesn't really say anything that gives away its provenance
19:09:20 <DanC_> I find MarcA's 1/23/93 announcement of Mosaic 0.5
(cc'd www-talk; are the archives complete in this respect?)
19:10:23 <sbp> Thomas R. Bruce refers to BLOCKQUOTE as "nebulous,
contemplated" on 12 Apr 1993
19:12:46 <sbp> "<blockquote> now supported." says Marc on 4 Jun 1993,
of Mosaic 1.1
19:13:34 <ww> DanC_: your email "HTML todo list" is missing #3...
19:13:41 <DanC_> are you using
to browse?
19:13:45 <sbp> ww: it was always insensitive, I think, and then the
choice was made in XHTML that lowercase should be used. this bothered
me at the time because I used uppercase. I don't know why they made
that decision
19:14:00 <sbp> DanC_: nope, I'm using the contents of
19:14:23 <DanC_> uppercase would have been a crime against humanity:
it would have cost more RSI. (shift keys are not free)
19:14:24 <sbp> then I Googled a substring to find where it was hosted
somewhere that I could link
19:14:45 <aindilis> caps
19:14:59 <sbp> also it would have meant various fixed attribute values
ought to be capitalised too
19:15:30 <sbp> (I prefer lowercase now, with the benefit of nearly two
decades of hindsight)
19:16:15 <ww> i always thought the contrast from differing case made
it easier to read... but maybe you're right...
19:16:50 <DanC_> ah... I find email exchanges with Dale Dougherty and
Terry Allen about docbook; maybe THAT's where I got blockquote?!
19:17:05 <sbp> I believe I already checked docbook. I'll look again
19:17:18 <DanC_> well, it certainly has it now:
19:17:41 <sbp> there's some early stuff in
19:18:02 <sbp> HMM!
has blockquote!
19:18:14 <DanC_> date?
19:18:28 <sbp> well, it's 14-Mar-1993 but then docbook1.0.dtd has the
same date so that's suspicious
19:18:54 <sbp> and docbook1.0.dtd has the copyright date 1992, and
also contains blockquote
19:19:04 <DanC_> bingo
19:19:12 <sbp> mystery solved!
19:19:26 <sbp> what was the nature of the thread with Dougherty and
Allen? does it mention blockquote?
19:19:58 <DanC_> "Span attribute of IndexTerm"
19:20:02 <DanC_> not so far...
19:20:48 <DanC_> well, there's  DocBook DTD 1.1 beta 1/19/93 from Terry Allen
19:20:54 <DanC_> with blockquote
19:21:00 <sbp> the logger here has gone bonk. is it okay if I package
the logs up from today and send them to www-archive or somewhere else
19:21:45 <sbp> (okay with you too, ww and aindilis? this place is
normally logged so I assume folks are comfortable, but just in case
anybody has got used to the logger not being here I figure I should
check. it hasn't been seen since February)
19:22:07 <DanC_> yes, though the logger is out to lunch, the social
contract remains in tact, at least for me
19:22:19 <DanC_> 12/4/92 is the 1st appearance in www-talk, right?
19:22:23 <sbp> yep
19:22:38 <sbp> in
19:22:54 <DanC_> I had been working with the docbook folks since 1991
(per )
19:23:02 <sbp> in fact it's the first appearance in an HTML context
overall that I've found so far
19:23:51 <DanC_> Terry's message was to ; he might have archives
19:24:16 <DanC_> or Con O'Connel. Or maybe even Norm Walsh, though he
came along much later, I think
19:24:35 <DanC_> ...
19:24:59 <DanC_> "Davenport Group FTP Archive If you have problems or
questions, please don't hesitate to contact Norman Walsh
19:25:09 <DanC_> yeah; I've got a problem; the archive has gone 404 :)
19:25:33 <DanC_> anyway... the social contract to publish was in place all along, evidently
19:25:44 <sbp> ah, I found
19:25:55 <sbp> which only contains docbook.dtd
19:26:21 <sbp> and is dated 12 Nov 1992 according to the zip metadata
for the file
19:27:22 <sbp> (1.1 is dated 26 Nov 1996, so that's been touched for
some reason)
19:27:55 <DanC_> 12 Nov 1992 <  12/4/92
19:28:04 <DanC_> anyway... the evidence seems to be piling up
19:29:19 <DanC_> grr...
only goes back to 199910
19:29:47 <DanC_> so... 2009... we solved the riddle in under a decade!
19:29:48 <DanC_> ;-)
19:31:37 <DanC_> do we have my  "HTML DTD issues"  msg of 11/19/92 to www-talk?
19:31:38 <DanC_> 5. What about <HP1> thru <HP5>... should we include them?
19:31:38 <DanC_> I'd prefer <em>, <tt>, <cite>, ala TeX. Or we could
19:31:38 <DanC_> go with the O'Reilly/Hal DocBook tags:
19:31:38 <DanC_> <Emphasis>, <OopsChar>, <wordasword>,<CiteBook>,<Subscript>,
19:31:38 <DanC_> <Superscript>.
19:32:49 <DanC_> yes...
19:33:10 <DanC_> that establishes that docbook tags were candidates for HTML
19:35:31 <DanC_> another shar archive 11/23/92 ... this time to TimBL
for the nextstep browser.
19:35:41 <DanC_> remind me to publish that one
19:40:13 <sbp> ah, that's interesting. and it's a week after the first
docbook DTD
19:40:24 <sbp> did you inspire them to write and maintain a DTD, as
you did with HTML?
19:41:57 <sbp> the first HTML DTD was written on 2 Jun 1992, and
documented in
19:43:37 <DanC_> no, the docbook folks were SGML before I was

Sean B. Palmer,

Received on Monday, 28 August 2017 19:57:02 UTC