- Wednesday 31st July
Moved the "entries submitted" line on the
home page of the virtual lottery below the group of links
because I finally decided that it was confusing to have the closing date
and "entries submitted up until..." date on consecutive lines. Hey, only
took me 9 months to spot this one ! Did the usual spell checking (picked up
some oldies I'd missed !), re-compilation with gcc, weblinting, linting and
internal link checking.
- Tuesday 30th July
Improved the checking for Netscape clients when issuing a warning about
not saving or printing the returned page from a virtual lottery entry or
lottery perms selection. I now check the agent string for "Mozilla/0*"
(Netscape 0.9), "Mozilla/1*" (Netscape 1.X) or "Mozilla/2*b*" (Netscape
2.0betas, which should all have expired by now anyway, but some people might
set their clocks back !) but not "MSIE" (Internet Explorer currently
fakes Mozilla/1.22 as its browser !!) before issuing the warning. In other
words, only truly pre-2.0 final Netscape users should be warned.
I also added the draw date to the returned page from a virtual lottery
submission - I had the virtual lottery number, but not the date of the draw
and someone noticed this recently and suggested I put it in. Quite a while
ago, someone suggested "credit card style" splits of the 16 hex digits of
the security code into 4 groups of four hex digits. This may be something I'll
consider in the future, although browser line wrapping might make a mess of
Changed the "Odds of Winning" page to be a "Chances of Winning"
page instead so that the figures would be
"1 in X" rather than the incorrect "X to 1" (should have been "X-1 to 1" of
- Monday 29th July
Got into work and found out that our Departmental network was basically
dead because of a network switch fault, which wasn't resolved until about
2.30pm. (Finally) uploaded the results for
Lottery #89. Moved the second teletext e-mail
request back 5 minutes to 8.20pm. My home phone was fixed today...phew.
- Sunday 28th July
Picked up the phone at home to contact Camelot for results confirmation,
but it was dead ! BT will send an engineer out tomorrow to fix it
- Saturday 27th July
The first teletext e-mail request came back at 8.05pm, but wasn't much use
(just ??'s) and the second request returned at 8.15pm with a clone version of
the first request. Chris Prickett's site had the results at 8.15pm, but no-one
else did, so I ignored the WWW results. The BBC 2 teletext mail request
came through at 8.35pm and the pages duly updated with the winning numbers
and a £9.6m jackpot forecast. The exact jackpot came through at 10.35pm.
- Friday 26th July
The Camelot fax finally arrived (10.08am) and it fixed an incorrect
Lucky Dip figure given to me on the phone yesterday...grrr.
- Thursday 25th July
Camelot were slow with the fax (didn't arrive today either), so I phoned them
up and only half the mid-week results were available - no Good Causes or
unclaimed prizes figures yet. :-(
- Wednesday 24th July
Slightly improved the code (again !) which flags some sections as "updated" on
the home page. This time, if no sub-page contains a "new" or "updated" icon,
then the code now finds the oldest .bhs file as its reference to determine if
the home page needs to flag the section as updated or not (previously, it
would use a random file from that section !). Since the oldest such file
is usually several months old, this should avoid the appearance of an
"updated" icon when a non-index.bhs file in the sub-section had been edited
to remove a "new" or "updated" icon.
- Tuesday 23rd July
Adjusted some of the pages for improved line spacing when viewed with lynx
(a UNIX text-only browser in case you're wondering), but it's probably time to
upgrade to the latest version of that to see if they've fixed those annoying
line spacing problems that I've been working around for ages now. For example,
two consecutive <P> tags should be collapsed to one blank line,
but lynx displays two :-( There's also no line spacing around certain tags
(e.g. <UL>...</UL> and <PRE>...</PRE>), but all
graphical browsers I've seen do put a blank line either side of those
tags when rendering them.
- Monday 22nd July
Uploaded the results for
Lottery #88. Added year navigation in the table
and pre-formatted headers of the yearly winning numbers because there was no
way to go forwards or backwards a year without going back up to the
index page and back down again. Set CELLPADDING
and CELLSPACING to zero on the HTML tables, although Netscape 3.X doesn't seem
to render this correctly - it uses a setting of one and leaves a gap between
table cells...a definite bug that isn't present in Internet Explorer.
Netscape 3.X also doesn't properly align main table cells that contain
<BR> tags when the table header cells are narrower than their
corresponding main cells - it appears that the line break causes following
spaces to be ignored when calculating the overall alignment of the lines of
that cell. In other words, ALIGN=right honours a trailing space in a line (just
before the line break), but not the space after a line break at the start of
the next line...duh !
It would be nice if individual cells could be padded to get around this (i.e.
<TD CELLPADDING=3>), but Netscape 3.X doesn't understand this and I
suspect it may have been overlooked in HTML 3.2. :-(
I eventually found a workaround by using multiple "hard spaces" ( )
in the table header cells to force the main cells to be wider and hence make
the alignment actually work. Not a good solution, but at least Netscape 2.X,
3.X and Internet Explorer 3.X render the table with reasonable padding now.
Moved the second teletext mail request back to 8.15pm, since it still returned
a clone of the 8.00pm request.
Someone told me about an old version of the Draw Order page that
was still hanging around from a couple of months ago, so I've replaced it with
an "oh no, it's moved" page. However, I'm not telling you where that page is
since I don't want people to link to it...
- Saturday 20th July
The auto-update system was right on the button for once and updated the
pages with an estimated jackpot (£9.6m) at 8.06pm, quicker than anyone
else on the Internet it appears. An exact jackpot came through at 9.17pm,
which was pretty nippy too.
- Thursday 18th July
Belatedly received the weekly fax from Camelot with the mid-week sales
figures and duly typed those in. I also got a letter this week from
Camelot (arrived on Wednesday) saying that they'd send me a paper copy
of results via post as well, but [in addition to the usual prize amounts
and numbers of winners] it only had estimates for scratchcards
and Good Causes figures, which isn't much good, especially by Wednesday
when the exact figures are faxed to me anyway !
- Wednesday 17th July
Added a new routine that is more intelligent about when to switch on or off
the "updated" icons on the home page. For the Wheeling Challenge,
Background Info and What's New sections, it now scans for the latest bhs file
(excluding the index.bhs document of course) in those sections that has
at least one "new" or "updated" icon within the document. The age of that
latest file then determines if the section will be flagged as "updated" or not.
The reason for adding this new code was to circumvent the common problem of
a manually maintained index.bhs page (particularly for the Background Info
section) having an updated timestamp just because I removed a "new" or
"updated" icon from the index page ! Fixed the
GIF for the fourth cheque I scanned in, because
I accidentally pasted a bit of stray text in there (before you ask, no it
wasn't extra zeroes :-) ).
- Tuesday 16th July
I've continued the HTML table progress I'd made with the index page
of the winning numbers section by extending
the code to the individual winning number HTML tables as well. It involved
keeping a set of table columns in memory as I constructed them with
<BR> tags to separate lines, rather than individual <TD>
tags for each table item. I only output the table columns when either
the expiry boundary is reached or the code hits the end of the table.
I've removed the table border to see if that looks better and also reduced the
length of the expiry boundary message (to close the HTML table columns up).
I've trimmed the "Winning Numbers" or "Draw Order" label at the top of the table
to just span the winning number columns, which makes it look a bit neater
(well, I like the way it juts out from the top now, so there !). Finally, I
removed the old X mosaic message since I doubt even die-hards would run that
when you've got Netscape 3.X to play with under UNIX. All these table changes
reduced the length of most pages by about 50% would you believe, so if nothing
else, they should appear a lot more quickly now.
There's still more work to do, such as experimenting with
CELLPADDING and CELLSPACING attributes, which don't seem to work correctly
(especially when set to 0) with Netscape 3.X, but seem OK with Internet
The lowest, highest and average summaries at the bottom of the tables are
still in the "old" format (individual cells) and, of course, at some point, the
numerical analysis section needs HTML tables introducing
(probably with the year splits like the winning numbers as well).
- Monday 15th July
Uploaded the results for
Lottery #87. The official Camelot site had
6 of its 7 ball GIFs "broken" this morning (wrong .jpg file reference) and
this was compounded by the lack of ALT tags on their ball graphics, meaning
that you had to read the frame source code to find the winning numbers !
Nice one, Camelot. Adjusted the second teletext mail request back to 8.10pm.
- Saturday 13th July
Yet again, the first two teletext mail requests (8.00pm and 8.09pm) were too
and returned identical results (no numbers), which is a result of the stupid
teletext mail service (wrongly) caching close requests for the same page.
Chris Prickett's site updated at around 8.15pm, but no others did, so it wasn't
until an 8.38pm return from a BBC 2 teletext mail request appeared that these
pages were updated. At 9.14pm, an ITV teletext estimated jackpot of £21.9m
turned up and an exact jackpot pool figure finally arrived at 10.36pm.
- Friday 12th July
I received my winning cheque yesterday and it's now been scanned in and
put on the usual
- Thursday 11th July
best performing tickets code to allow
for both 0-match and 1-match [or more] to be excluded from the output (there's
only 2 ticket combinations that have failed to manage a 2-match, so any day
now they'll achieve this and the page will now auto-compensate appropriately).
I've also changed the code so it can display an arbitrary number of tickets
that have won the most prizes (currently set to 10). BTW, no
other lottery site on the Internet has an equivalent page yet (because the
code is quite complex and needs to be fine-tuned to run as fast as possible).
Camelot's weekly fax to me came through yesterday at 12.12pm, but I was so
busy, I didn't have the time to check for it until this morning. As usual,
I updated the mid-week sales figures now that I'm not running around like
a headless chicken :-)
- Wednesday 10th July
Removed any links to Might BU (my first real rival) on these
pages because they appeared to have packed it in now that my site and Chris
Prickett's site are providing overkill w.r.t. lottery facts'n'figures. It now
also means I'm down to 5 WWW sites I can scan for results (yes, there's other
3 UK lottery sites - including Camelot's - I don't scan, but this is because
they refuse to display the results as plain text, instead preferring
unparseable table code or graphics with no ALT tags).
- Tuesday 9th July
Bought two more tickets for next Saturday's rollover,
which brought my total spending on the lottery to within £1 of the magic
- Monday 8th July
A belated £20m jackpot pool estimate came through via teletext e-mail at
7.10am. Uploaded the results from
Lottery #86. I
won another £10 prize (my fifth overall),
but I still remain nearly £150 down of course. I discovered that
Camelot had finally launched their new WWW
site, although it isn't up to much
(fancy frames design and some average background info, but zero in the way
I've put the second teletext e-mail request back a minute to 8.09pm.
InterLotto changed their background graphic and text/link colours recently,
so I've changed my
stats area to match them.
- Saturday 6th July
As usual, the first couple of e-mail requests were too fast (they just got
the first number), so the backup WWW scan found Mersinet's update first
(around 8.20pm) and Chris Prickett's site second (around 8.50pm). This
second grab prompted an update to these pages at 8.50pm as well. The first
teletext e-mail with a results update came through at 9.14pm and this
changed the estimated jackpot to £11.1m.
The exact rollover figure
came through at 10.33pm, although shortly after this, the Computing Services
Department switched off their teletext mail service and there were no more
updates until Monday morning.
- Friday 5th July
I started a revamp of the way I generate the HTML tables for the winning
numbers. Instead of individual <TD> tags for each and every bit of
data, I've grouped together each coloured section into a single table cell and
used <BR> to break to the next line in that cell. It seemed to work
well, so I've adjusted the winning numbers
index page this way and will tackle the main
winning number tables over the next few weeks. The other benefit is that the
cell tag verbiage is significantly reduced and the new tables will load quicker
! The only downside is that 2.7beta releases of X mosaic mush the table into
nothing because they refuse to honour the <BR> tags inside a cell...grrr.
Received a letter from Camelot about my query as to what they'll be doing on
the Internet. They are setting up a new WWW site,
"http://www.national-lottery.co.uk/", which they will launch sometime this
month and will "develop it over the next 12 months". I guess this means that
it will initially just have winning numbers, prize amounts and number of
winners with little or no analysis.
Of course, it provides another external WWW site to grab the results from
(and maybe also the number of winners/prize amounts if that's consistently
presented), but it does have a downside: it'll pull people away from my
site simply because it's the official one. I wonder if they'll farm the
site design out to some dubious Web design company who'll go overboard with
frames and fancy graphics ? I suspect the initial site will be all style
and little content...
- Thursday 4th July
Received the Internet Now
Entertainment Edition CD-ROM free through
the post because a March 1996 copy of my lottery pages had been included
(with my permission) by
Instant Access. What worries me is that
they are selling a paperback book and CD-ROM for a total of £9.99, which
is little steep compared to equivalent Internet mags+CD-ROMs at around
- Wednesday 3rd July
Camelot have promised to fax me each Wednesday morning from now on with the
mid-week figures (scratchcards, unclaimed prizes etc.) ! I received the first
fax this morning at 10.47am and duly updated the pages with the info.
- Tuesday 2nd July
Corrected a mis-alignment in the grand total line for the 3-match
total prize pools. Adjusted the WWW
grab config file for Mark Butler's results site because he changed the layout
a few weeks back without me noticing. Also disabled grabbing Might BU
because they've temporarily pulled their pages.
Interestingly, there were three successful WWW page grabs and three failures
over the weekend. The Lotto Fun Club simply hadn't updated for the 22nd
June results, never mind the 29th June (if a site doesn't do this, I just
ignore them completely for the rest of the weekend). Next up, Chris
Prickett forgot to put an ALT tag with his page update, so there was no
way to determine the winning numbers non-graphically. Finally, I've already
mentioned the mis-parsing of Mark Butler's site because of the (slight) change
in format a few weeks ago.
There was a discrepancy between the way the
triples pages were
calculated and the corresponding pairs/triples figures on the
trivia analysis page. In the pairs and triples
pages, I truly just found pairs or triples of numbers regardless of their
neighbouring numbers. However, in the Pairs and Triples columns in the trivia
analysis, I excluded pairs that were part of triples or, indeed, triples
that were part of quads. I decided to drop this exclusion, which not only
simplified the code involved, but also synchronised that fact that a pair
can be part of a triple or a triple part of a quad on both sets of pages.
- Monday 1st July
Teletext e-mail finally came through for the first time at 8.35am, so the
exact jackpot amount was belatedly updated.
Uploaded the results for
Fixed a bug with the What's New link on the home page - it was advancing
to the next month's What's New page earlier than it should have done.
Several e-mail replies strangely came through at 3.17pm from the weekend's
teletext requests (i.e. the teletext mail server had processed some this
morning and then the rest in the afternoon). I spotted this and ended up
having to re-generate the pages to remove the unwanted "update".
Fixed a minor bug in the
lottery links generation code which was duplicating the
"(New)" tag several times next to a new lottery site.