Re: Claim by 'Stevo' - "CEOP show Maddie is missing on 30th April 2007"

by Guest 21.06.15 12:50

[quote="BlueBag"]I think WBM [u]could[/u] be used in legal cases... IF... the archived page contains it's own time stamp.

Like many BBC pages do, and newspapers... Date and time of article.

Or forum posts.

I think after what we've seen here a good lawyer could wriggle out of something based on the WBM timestamp.[/quote]

============================

Any lawyer

Bang goes WBM's unblemished record

If I were CB, I'd be on that bike to Scotland with KH

by Guest 21.06.15 12:54

aiyoyo wrote:
All past court cases that relied on their info for successful convictions are now brought into question.

Not if date and time were embedded in the page contents itself. The WBM timestamp becomes secondary.

by worriedmum 21.06.15 12:57

Coming to this (rather bewildering) topic a bit late, I did wonder if it is possible that the date 30 April covers the earliest parameter of the screen grab and therefore anything between that date and the end parameter date is included. So, it doesn't actually mean it was written on that date?

by HKP 21.06.15 12:58

@BB not sure I know what you mean. There's not a file in that folder for October(as far as we know) but the controversy is over links on the page to October stories. As far as can be ascertained a crawl was made on 30/04 and it found what it found, its then a case of whether some data 'refreshes' at a future date, the FAQs say it does.

by Guest 21.06.15 13:16

HKP wrote:@BB not sure I know what you mean. There's not a file in that folder for October(as far as we know) but the controversy is over links on the page to October stories. As far as can be ascertained a crawl was made on 30/04 and it found what it found, its then a case of whether some data 'refreshes' at a future date, the FAQs say it does.

None of the following are redirects:

[You must be registered and logged in to see this link.]
[You must be registered and logged in to see this link.]

[You must be registered and logged in to see this link.]

That's the same page archived on three different days in three different WBM folders.

The URL remains the same when you open the page.

If there WAS a folder and file:
[You must be registered and logged in to see this link.]

..that had a page with October data news release on it that would convince me that the WBM was broken in October and put the the October pages in a april folder.

My question is for anyone who was in before the pages were deleted, is this the case?

by LG 21.06.15 13:36

@ BlueBag - sorry that was my mistake - I did post both (one above the other) way back (no pun intended) on this thread.

Whilst here, does anyone know of another archive other than Way Back?

[You must be registered and logged in to see this image.]

by sar 21.06.15 13:40

BlueBag wrote:
whatsupdoc wrote:
cloak'ndagger wrote:I can vouch for Syn.. .she may not be saying what we want to hear but she is an expert in the technological field.

I'll second that. I know of syn0nymph from Twitter.

I don't think we should be knocking newcomers either.
History shows we should be highly suspicious of new comers.

Especially hot single issue ones come to "correct" us.

+1 BB

by sar 21.06.15 13:41

whatsupdoc wrote:Meanwhile, back at the ranch, I've been checking the mccann.html files on WBM and if you try 30th April and the 1st May links including an * after the date/time stamp you get the calendar and can see the 30th and the 1st snapshot dates, Clicking on either will bring up a pic of Madeleine but the date on the top has been changed to 13th May.

The link for YouTube around midday on the 30th April is still there and showing 30th April when the snapshot is clicked on. All working, no problems.

So other sites such as YouTube are working on 30th April around 1200 on WBM ...just the mccann.html file causing a problem....AFAIK

I don't think a glitch covers an event such as making a file entry days before it has supposedly occured.

+1 whatsupdoc

by Doug D 21.06.15 13:53

BlueBag,

Don’t know if LG’s latest picture helps you, but a link to HideHo’s screengrabs here:

[You must be registered and logged in to see this link.]

However, I suspect you want the http line above what this shows.

Don’t know if someone can contact Steve Marsden on FB to see if he has got fuller screenshots.

………………………………………………………

Not new information as such and don’t know if it’s of any use to anyone but from the WBM CEOP capture of 27/4/07 the forward blue arrow link has been broken, although the whooshed capture of 30^th still flashes up. The page just gets stuck on 27^th. Trying to go both ways the ‘http’ line flashes for the 30^th but then just reverts.

[You must be registered and logged in to see this link.]

Coming back the other way from 14/5/07 doesn’t seem to be a problem. The 30/4 date still flashes up, but just takes you to 27^th April.

Similarly with the ‘index’ site:

[You must be registered and logged in to see this link.]

the back arrow flashes up with the 30/4 date, the ‘https:’ line at the top looks for the 30/4 then drops back to 12^th May and won’t work going backwards, just drops you back to 12/5.
Going forwards, the 8^th April is there, the 30^th April flashes on the blue forward arrow, it looks for 30^th April on https: line , then jumps to 12^th May.

We know the entries for the 30^th have been whooshed, but is there any reason for the breaks in these arrow paths, other than a poor whooshing job at WBM?

by Guest 21.06.15 13:53

LG wrote:@ BlueBag - sorry that was my mistake - I did post both (one above the other) way back (no pun intended) on this thread.

Whilst here, does anyone know of another archive other than Way Back?

[You must be registered and logged in to see this image.]

OK that does it for me.

I think WMB was broken in October 2007 and put October pages in an April folder.

The most likely explanation is that the mccan.html was saved in October in the same process as the above... specifically in a process that gave the timestamp as 2007/04/30 11:58:03 - all the CEOP pages have this timestamp.

And as a programmer of 34 years standing that doesn't really surprise me.

Although I would like to see other examples because there must be some.

Now.. why was I so hard on Nuala and syn? Because they were clouding the issue with rubbish.

by notlongnow 21.06.15 14:01

BlueBag wrote:
LG wrote:@ BlueBag - sorry that was my mistake - I did post both (one above the other) way back (no pun intended) on this thread.

Whilst here, does anyone know of another archive other than Way Back?

[You must be registered and logged in to see this image.]
OK that does it for me.

I think WMB was broken in October 2007 and put October pages in an April folder.

The most likely explanation is that the mccan.html was saved in October in the same process as the above... specifically in a process that gave the timestamp as 2007/04/30 11:58:03 - all the CEOP pages have this timestamp.

And as a programmer of 34 years standing that doesn't really surprise me.

Although I would like to see other examples because there must be some.

Now.. why was I so hard on Nuala and syn? Because they were clouding the issue with rubbish.

How unfortunate if true for wbm,they couldn't have have picked a worse day to for it to be in.

by HKP 21.06.15 14:11

@notolngnow, exactly they couldn't have picked a worse day if they tried. It seems pretty clear to me that there was a file somewhere either waiting to be uploaded or not quite fully completed and Wayback found it. The back tracking is just one of the red flags. Somehow when another crawl has been done the original has kept this data (dynamic) all IMO.

@Tony I agree (about this topic)
@LIR thanks

by rustyjames 21.06.15 14:27

I wish we could just concentrate on the technical discussion. For the record I agree with the vast bulk of BlueBag, Nuala, syn and Richard Hall's posts, but there are elements of them I disagree with; however, angry and personal exchanges aren't going to help.

The attachements to the email below are the first thing we could really do with, however I suspect even those are not going to be provide enough information given the way the page is generated. All we are seeing is the view that the Query and Replay UI's are presenting. The source information in the Resource Store is what is of real interest but only the Internet Archive has that. (Lots more technical info here - [You must be registered and logged in to see this link.]).

[You must be registered and logged in to see this image.]

by Guest 21.06.15 14:40

rustyjames wrote:I wish we could just concentrate on the technical discussion. For the record I agree with the vast bulk of BlueBag, Nuala, syn and Richard Hall's posts, but there are elements of them I disagree with; however, angry and personal exchanges aren't going to help.

The attachements to the email below are the first thing we could really do with, however I suspect even those are not going to be provide enough information given the way the page is generated. All we are seeing is the view that the Query and Replay UI's are presenting. The source information in the Resource Store is what is of real interest but only the Internet Archive has that. (Lots more technical info here - [You must be registered and logged in to see this link.]).

[You must be registered and logged in to see this image.]

Yes.. this was before Chris Butler realised they had a problem.

There is no question that at least one October CEOP page (the home page) was in the April 30th 2007 folder.

And seeing as they all have the same archive date and time - to the second - you have assume the likelihood they were all from October 2007 trawl.

by HKP 21.06.15 14:46

@Rusty, did that document you highlighted tell us anything worthwhile (excuse my ignorance as I didn't really see anything)?

by Wahrheit 21.06.15 15:15

BlueBag wrote:Yes.. this was before Chris Butler realised they had a problem.

There is no question that at least one October CEOP page (the home page) was in the April 30th 2007 folder.

And seeing as they all have the same archive date and time - to the second - you have assume the likelihood they were all from October 2007 trawl.

Hi,

It is a possibility but I think it is not time to concede that it is the most probable reason yet.

For that to happen the error would need to be satisfactorily explained and/or an example (or several) of a strictly analogous error would have to be identified.

The question of whether it is possible for dynamic data from a future date to enter a reconstituted/replayed page would have to be debunked definitively (it hasn't been so far). From my reading that is currently the more probable scenario at this point given that WBM itself points out this possibility as a "feature" (if I can use the term) of their system.

I don't "mind" one way or the other how the discussion turns out but I am as interested as the next person to understand what's behind this.

by whodunnit 21.06.15 15:26

Again, the WMB FAQ says links will automatically update to the next available update.

CEOP site doesn't seem to update it's 'latest news' very often so October may have been the next available update to the link provided in the sidebar. [Script can be written that automatically converts links to text.]

For an example of updated links on an archived page go to page 47 of this thread and see the link to the April 30, 2007 archived CNN page provided by member SixMillionQuid. Click on any link therein. Some of the links were never updated but some were and redirect to 'future' dates.

by Guest 21.06.15 15:26

Wahrheit wrote:

The question of whether it is possible for dynamic data from a future date to enter a reconstituted/replayed page would have to be debunked definitively (it hasn't been so far). From my reading that is currently the more probable scenario at this point given that WBM itself points out this possibility as a "feature" (if I can use the term) of their system.

Could you explain that in different terms?

by rustyjames 21.06.15 15:31

HKP wrote:@Rusty, did that document you highlighted tell us anything worthwhile (excuse my ignorance as I didn't really see anything)?

Useful to help understand and confirm the architecture - now know far more about web archival than I ever wanted :-)

Googling about the web archival files has thrown up lots more info, but I simply don't have time to read it all. I did find this interesting article that explains how content for a page is actually scattered across lots of random archival files - [You must be registered and logged in to see this link.]

The crawled info seems to be in an arc (or more recently warc) file which contains not only what was found but lots of meta-data such as timestamps, ip addresses etc. These form the Resource Store.

The Resource Index as the name suggests indexes the content for efficient retrieval.

The Query UI again as the name suggests parses the request for particular pages / date ranges etc and presents the results (e.g. the calendar page).

The Replay UI is what uses the above information to render a page and alter links etc so that the appropriate content is displayed. In 2007 the URL rewriting was done on the client side using JavaScript; it now seems to be server side. That's irrelevant though as it's just a change in presentation.

Again what is of most interest is what is in those archive files in the Resource Store as it is those I'd trust the most as the single source of truth and the metadata could be interesting.

by Wahrheit 21.06.15 15:32

BlueBag wrote:
Wahrheit wrote:

The question of whether it is possible for dynamic data from a future date to enter a reconstituted/replayed page would have to be debunked definitively (it hasn't been so far). From my reading that is currently the more probable scenario at this point given that WBM itself points out this possibility as a "feature" (if I can use the term) of their system.
Could you explain that in different terms?

I think whodunnit has explained what I meant in the previous post.

by Guest 21.06.15 15:33

whodunnit wrote:Again, the WMB FAQ says links will automatically update to the next available update.

CEOP site doesn't seem to update it's 'latest news' very often so October may have been the next available update to the link provided in the sidebar. [Script can be written that automatically converts links to text.]

For an example of updated links on an archived page go to page 47 of this thread and see the link to the April 30, 2007 archived CNN page provided by member SixMillionQuid. Click on any link therein. Some of the links were never updated but some were and redirect to 'future' dates.

The home page.. the base CEOP page for October ended up in an April folder... [You must be registered and logged in to see this link.] some latest news threads were on that page but it wasn't the latest news page.

We have a hard home CEOP october web page in an April folder.

That suggests to me that happened no earlier than october.

And because of the identical timestamp - to the second - it suggests the other pages with the same timestamp were trawled with it.

by Guest 21.06.15 15:34

Wahrheit wrote:
BlueBag wrote:
Wahrheit wrote:

The question of whether it is possible for dynamic data from a future date to enter a reconstituted/replayed page would have to be debunked definitively (it hasn't been so far). From my reading that is currently the more probable scenario at this point given that WBM itself points out this possibility as a "feature" (if I can use the term) of their system.
Could you explain that in different terms?
I think whodunnit has explained what I meant in the previous post.

I understand the linked pages thing.

This is a homepage.

by whodunnit 21.06.15 15:52

BlueBag Today at 3:33 pm

whodunnit wrote:Again, the WMB FAQ says links will automatically update to the next available update.

CEOP site doesn't seem to update it's 'latest news' very often so October may have been the next available update to the link provided in the sidebar. [Script can be written that automatically converts links to text.]

For an example of updated links on an archived page go to page 47 of this thread and see the link to the April 30, 2007 archived CNN page provided by member SixMillionQuid. Click on any link therein. Some of the links were never updated but some were and redirect to 'future' dates.

The home page.. the base CEOP page for October ended up in an April folder..[link to ceop archive removed] some latest news threads were on that page but it wasn't the latest news page.

~~~~

Then it was also broken in July when it put a poster created in July into the mccann.html file which was originally captured on April 30, 2007 but which now redirects to May 13.

Boy, that sure is a lot of broke which ONLY effects pages having to do with the Mccann case. The most logical explanation is the links updated. Now the October links on the home page are 'fixed'.

by Joss 21.06.15 15:56

What was so significant on April 30th. i wonder? for that date to even show up?

by Guest 21.06.15 16:08

whodunnit wrote:
Then it was also broken in July when it put a poster created in July into the mccann.html file which was originally captured on April 30, 2007 but which now redirects to May 13.

The poster may well have been created in July but it seems it was put in the April folder in October at exactly the same date and time as the homepage - the October homepage was definitely created in October.

Boy, that sure is a lot of broke which ONLY effects pages having to do with the Mccann case. The most logical explanation is the links updated. Now the October links on the home page are 'fixed'.

We don't know if it only affects this web site. I now expect many others to have the same problem. I WOULD like to see examples.

Look.... before anyone dug too deep and before any questions were asked.. before WBM knew they had a problem.. someone found the October homepage in the April 30th 11:58:03 folder and well as the mcann.html page. That is a fact. There are screenshots.

They've deleted the whole folder since.... that was their response to the problem being highlighted.

They might have dug around themselves and deleted other bad folders for other sites too by now.

And they may never admit it.

After the index rebuild the overview page will lose the April 30 links and any others they found which we will never know about.

by whodunnit 21.06.15 16:11

I'm just going to partially quote WLBTS: "...it doesn't make sense to me that crawls done at various different times were all in error and yet managed to be put in the same timestamp folder."

by **Tony Bennett** 21.06.15 16:28

Joss wrote:What was so significant on April 30th, I wonder? - for that date to even show up?

You could try asking Robert Murat. This was the day someone 'phoned him up and said 'Come over to Praia da Luz - quick!'

And he did just that, leaving home at about 5.00am the following morning to catch the 7.00am flight from Exeter to Faro.

And then he lied comprehensively to the police about what he did when he got over there

by rustyjames 21.06.15 16:48

Just been playing around with the Wayback/Internet Archive's API's.

Haven't got anywhere to store the output online, but this link selects all index entries for the CEOP site on April 30th 2007 - [You must be registered and logged in to see this link.]

There are entries in there for news articles in the future.

by Joss 21.06.15 16:56

rustyjames wrote:Just been playing around with the Wayback/Internet Archive's API's.

Haven't got anywhere to store the output online, but this link selects all index entries for the CEOP site on April 30th 2007 - [You must be registered and logged in to see this link.]

There are entries in there for news articles in the future.

So what does that mean?

by rustyjames 21.06.15 17:20

Joss wrote:
rustyjames wrote:Just been playing around with the Wayback/Internet Archive's API's.

Haven't got anywhere to store the output online, but this link selects all index entries for the CEOP site on April 30th 2007 - [You must be registered and logged in to see this link.]

There are entries in there for news articles in the future.
So what does that mean?

It means it's very strange that the following entries for example would be indexed on the 30th April

Code:: uk,gov,ceop)/(s(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070514_ceop.htm 20070430115803 http://www.ceop.gov.uk/(S(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070514_ceop.htm text/html 200 AETMTGNS3RN7LVV425NBZCGQNZS2UNVH 4568 uk,gov,ceop)/(s(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070607_ceop.htm 20070430115803 http://www.ceop.gov.uk/(S(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070607_ceop.htm text/html 200 3RYSYZYZJLC7INV6I6YOTQOE7CYSGL6G 3637 uk,gov,ceop)/(s(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070611_ceop.htm 20070430115803 http://www.ceop.gov.uk/(S(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070611_ceop.htm text/html 200 4IWXDFCOSAXWAJZVJI5EY7LD45AWDPJD 5797 uk,gov,ceop)/(s(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070613_ceop.htm 20070430115803 http://www.ceop.gov.uk/(S(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070613_ceop.htm text/html 200 FDPG6RPJ75WIRMOULKKBZ7CMTZO2VWO7 2605 uk,gov,ceop)/(s(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070625_ceop.htm 20070430115803 http://www.ceop.gov.uk/(S(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070625_ceop.htm text/html 200 JTVX33O6UVI4QMOLOXI6D7KMFLF3VLYZ 5159 uk,gov,ceop)/(s(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070719_ceop.htm 20070430115803 http://www.ceop.gov.uk/(S(0p4dtw454tlgkwy543bmxwfy))/news_items/article_20070719_ceop.htm text/html 200 H64KMBMA26EZOU7W4BCGQVQMHOJKHVBE 4165

Picking the 7th June article and searching for all references to it ([You must be registered and logged in to see this link.]) shows:

Code:: uk,gov,ceop)/news_items/article_20070607_ceop.htm 20070430115803 http://www.ceop.gov.uk/news_items/article_20070607_ceop.htm text/html 200 3RYSYZYZJLC7INV6I6YOTQOE7CYSGL6G 3609 uk,gov,ceop)/news_items/article_20070607_ceop.htm 20070708220101 http://www.ceop.gov.uk/news_items/article_20070607_ceop.htm text/html 200 W4SJBDOIS5X4EBLZ3KRWFDOW7MCLA23B 3511 uk,gov,ceop)/news_items/article_20070607_ceop.htm 20071230181833 http://www.ceop.gov.uk:80/news_items/article_20070607_ceop.htm text/html 200 3RYSYZYZJLC7INV6I6YOTQOE7CYSGL6G 3559 uk,gov,ceop)/news_items/article_20070607_ceop.htm 20080229143339 http://www.ceop.gov.uk:80/news_items/article_20070607_ceop.htm text/html 404 6IFP5M3IZ6HTG6U6BTIMGIDIUFS7R6NR 1919 uk,gov,ceop)/news_items/article_20070607_ceop.htm 20080501100235 http://www.ceop.gov.uk:80/news_items/article_20070607_ceop.htm text/html 404 6IFP5M3IZ6HTG6U6BTIMGIDIUFS7R6NR 1917

i.e. other than the 30th, all other references appear in indexing after the article was published, (the last couple have a status of 404 which means the page wasn't found). Other articles seem to be the same, for example after an April 30th entry, the 19th July article is then only next indexed on the 27th August.