The frame around these articles — the block column, the block names, the main menu, the video library, the login box, the rating, the print link — is qadsiyah.com’s own. It is read out of twenty-nine article pages that the Internet Archive holds from 17 February 2010, saved as PHP-Nuke served them. Every word in the furniture is copied from those files.
The page is <html dir="rtl">, and the body table’s row has
three cells: a column of blocks, the story, a second column of blocks. Because
the page is right-to-left the first cell is the right column and the third is the
left one. This is settled by byte position in the saved markup, not by eye: the
five blocks
دخول المنتدى، القائمة الرئيسية، مكتبة الفيديو، من يتصفح الآن، اللغات
all fall inside the first cell, the story inside the second, and the four blocks
دخول، روابط ذات صلة، تقييم المقال، خيارات inside the third.
A correction. The first version of this page put those last four blocks
underneath the story and said there had never been a left column. That was wrong.
Khalid produced the archived page and the left column is plainly there; the positions
above confirm it. Each block is built the same way: a cap, then a body on
#FEF8C1 whose heading cell is styled
border: 1px dotted #F9CA0D and whose heading text is
<font size="4" color="#B80C07">.
Not one theme picture. Across all twenty-nine saves the accompanying folders hold
six distinct files, and every one of them belongs to the Wayback toolbar. The
reason is on the site itself: qadsiyah.com’s robots.txt of 2007
forbade crawlers from /images/ and /themes/, and the archive
obeyed it. So these were never savable and are not here:
qs2_01, qs2_03, qs2_05,
qs2_09, qs2_11, and the 7px side strips
qs2_06 / qs2_08qadsiyah_01, qadsiyah_04,
qadsiyah_07b1gg.swf (replaced by the banner artwork itself, see below)stars-1.gif … stars-5.gif,
and print.gifDrawn in CSS, in the colours the surviving markup states for itself — not colours anyone picked in 2026:
| colour | where it is written | what it does here |
|---|---|---|
#FEF8C1 |
bgcolor on the block table | the block body |
#F9CA0D |
1px dotted on the heading cell | the block borders |
#B80C07 |
<font color> on the heading | headings, edges, caps |
#FFF7DD |
bgcolor on <body> | the page |
#FEF7BA |
bgcolor on the layout table | the wrapper |
#363636 |
link / vlink on <body> | links |
The banner. b1gg.swf is gone and no copy of it exists. What stands
in its place is not a guess and not a collage: it is the single 900×300 banner Khalid
found, whole, unretouched and uncut — crest in the middle, the supporter with the
ball at one corner, the man in the red shirt at the other, and the line
أول فريق سعودي يحرز كأس الكؤوس الآسيوية across the top. An earlier
version of this page cut that picture into three and laid the pieces over a widened,
blurred copy of its own background. Khalid said the rectangle it produced over the crest
was not close, and he was right; it is gone, and with it the possibility of any seam,
because there is now nothing to seam. The band is held to the width of the red-bordered
wrapper below so its edges line up with the page.
The movement is borrowed from 3badinet101.swf, an early-2000s forum banner
Khalid supplied so the idiom could be read off a real file rather than guessed at: one
still photograph, 150 frames at 12 frames a second — twelve and a half seconds —
with light drifting across it. So the picture fades up once and a soft light crosses it on
that same 12.5-second cycle. It is CSS, not Flash, which is the point: Ruffle is the one
thing on this site that can fail, and the header is the last place that should happen.
The white haze over the picture was cut by a fifth, at Khalid’s instruction, because
it was closing over the Arabic line. Anyone who has asked their browser not to animate
gets the finished frame and no sweep. If a copy of b1gg.swf ever turns up, it
drops straight into the same slot.
The Internet Archive’s own index lists 39 article pages for
qadsiyah.com: sid=4 through sid=43, with
sid=35 absent and nothing above 43. A second query for any other
address carrying a sid — print pages and the rest —
came back empty, so that list is the whole of it.
This section holds 38 of the 39. The one not here is
sid=22, and it is not a loss: Khalid, who published them, says 22
and 21 — نجران يجندل القادسية ويدفنه بالأخدود
— are the same report posted twice, and 21 is here. sid=35
was never captured by anyone. So every Al-Qadsiah article that survives, in
substance, is on these pages.
Khalid supplied 150 photographs out of his own folders. 112 stand in the
exact place the article asked for them, matched on the filename the page itself
names — an exact key, never a resemblance. 38 answer no slot at all
— either the article was published with no picture tags, or the file is one
he has and the page never carried — and those are shown in a gallery under
the story rather than discarded. 91 slots are still named and still empty:
those pictures were on qadsiyah.com’s server, the 2007 robots.txt
kept the crawler out of /images/, and nobody has them. Each keeps a box
carrying its own name.
Every one of the 150 is the original file, byte for byte — same bytes, same EXIF, same timestamp. An earlier pass re-encoded all of them through a JPEG save at quality 82; 141 were already small enough to need no such thing, and the round trip stripped the camera block off 120 of them. That block is evidence — it is how we know these were shot on an Olympus SP570UZ on the evening of 21 October 2008 — and a reconstruction should not be the thing that destroys it. Checksums confirm all 150 match their sources exactly.
The camera clock cannot place a renamed photograph. Khalid had renamed nineteen pictures after the people in them, which made them unplaceable by name, and the obvious idea was to use the time each was taken to work out where in the sequence it belonged. Eight photographs whose slot is already known still carry a timestamp, so each was hidden in turn and its slot predicted from the other seven. One came back right and five came back wrong. The method fails on cases where the answer is known, so it was not used on cases where it is not.
The captions were never lost, because they were never his. When Khalid
supplied the same nineteen under their real names, the descriptive names looked
like information about to be thrown away, so the two sets were matched on the
pictures themselves — a difference hash, which survives re-encoding. Seventeen
photographs appear in both sets under the same name and served as controls: all
seventeen matched themselves at distance 0, the nearest rival 17 bits away. The
nineteen then matched at distance 0 too, and h6.jpg resolved to
hf6.jpg, which is what Khalid had said it was. But set against the
article’s own text, seventeen of the eighteen names are the article’s own
caption line, word for word or all but, and the eighteenth is a shortened form
of one. He had named the files from the captions. Printing them would print every
name twice, so none is printed. The names were in the article the whole time.
The login box, the rating and the print link are shown as they stood, because they were part of the page a reader saw. None of them works: there is no server behind them any more. Menu entries lead to a notice naming the address they had. One link is real — دخول المنتدى opens the 2009 forum, which we do hold.
في صندوق «من يتصفح الآن» يظهر: 4 ضيوف وعضو واحد. هذا ما كان لحظة حفظ الصفحة في 17 فبراير 2010، وليس عدداً الآن — وهو مكتوب تحت الصندوق حتى لا يُقرأ على أنه حيّ.