msplinks.com

74 views
Skip to first unread message

Martin Atkins

unread,
Jul 7, 2008, 2:05:50 PM7/7/08
to social-g...@googlegroups.com

I noticed some time ago that the graph was getting messed up by
MySpace's obfuscated msplinks.com URLs, but they don't seem to be
showing up anymore.

Does the SGAPI parser now treat these msplinks.com URLs as their real
URL? Is it safe for me to put my rel="me" link back on my MySpace page?


Brad Fitzpatrick

unread,
Jul 7, 2008, 2:10:24 PM7/7/08
to social-g...@googlegroups.com
I know nothing of these links.  Could you elaborate?

Martin Atkins

unread,
Jul 7, 2008, 4:30:24 PM7/7/08
to social-g...@googlegroups.com
Brad Fitzpatrick wrote:
> I know nothing of these links. Could you elaborate?
>

MySpace obfuscates the URLs of all outgoing links in user-supplied
content on profile pages, by rewriting them to opaque nonsense in the
domain msplinks.com. It seems that they do this so that they can quickly
disable links to phishing sites or somesuch. However, this means that
when I do this:

<a href="http://martin.atkins.me.uk/" rel="me">

it actually becomes something like:

<a
href="http://www.msplinks.com/MDFodHRwOi8vbWFydGluLmF0a2lucy5tZS51ay8="
rel="me">

I guess there's already some cleverness going on with collapsing of
permanent redirects into a single node, but rather than an HTTP redirect
msplinks.com serves an intermediate page saying "You're leaving
MySpace!" or somesuch, with a link to the target URL that looks like this:

<a id="link" href="http://martin.atkins.me.uk/"></a>

There seems to be more about this here:
http://mashable.com/2007/04/24/msplinks/


Dale Newfield

unread,
Jul 7, 2008, 6:33:25 PM7/7/08
to social-g...@googlegroups.com
Martin Atkins wrote:
> when I do this:
>
> <a href="http://martin.atkins.me.uk/" rel="me">
>
> it actually becomes something like:
>
> <a
> href="http://www.msplinks.com/MDFodHRwOi8vbWFydGluLmF0a2lucy5tZS51ay8="
> rel="me">

This long string is simply a base64 encoding of the link in question.

MDFodHRwOi8vbWFydGluLmF0a2lucy5tZS51ay8=
is base64 for
01http://martin.atkins.me.uk/

I was able to verify this by generating a valid msplinks.com page
including a new link I encoded (after adding "01" to the beginning).

> I guess there's already some cleverness going on with collapsing of
> permanent redirects into a single node

Since base64 encoding is stable, all links to the same address will
encode to the same msplinks link.

-Dale

Martin Atkins

unread,
Jul 8, 2008, 2:36:28 AM7/8/08
to social-g...@googlegroups.com
Dale Newfield wrote:
> Martin Atkins wrote:
>> when I do this:
>>
>> <a href="http://martin.atkins.me.uk/" rel="me">
>>
>> it actually becomes something like:
>>
>> <a
>> href="http://www.msplinks.com/MDFodHRwOi8vbWFydGluLmF0a2lucy5tZS51ay8="
>> rel="me">
>
> This long string is simply a base64 encoding of the link in question.
>
> MDFodHRwOi8vbWFydGluLmF0a2lucy5tZS51ay8=
> is base64 for
> 01http://martin.atkins.me.uk/
>

Aha! I should've tried that. Thanks.

Should be easy for the SGAPI parser to process these, then, right Brad?

Brad Fitzpatrick

unread,
Jul 8, 2008, 2:48:48 AM7/8/08
to social-g...@googlegroups.com
Yeah, guess so.  Kinda ugly, but so is most the data on the web.  :-)

- Brad

Dale Newfield

unread,
May 11, 2010, 1:49:06 PM5/11/10
to social-g...@googlegroups.com
So, any chance that this nearly two year old request might get
implemented, so that I can wind up with a bidirectional rel="me" to a
myspace page, and thus be able to leverage the "friends" specified
there?

Specifically, on my myspace page http://www.myspace.com/dnewfield I
added <a href="http://newfield.org/dale" rel="me">, but of course they
replace that with <a
href="http://www.msplinks.com/MDFodHRwOi8vbmV3ZmllbGQub3JnL2RhbGU="
rel="me">, and while I could add a rel=me link from my real home page
to this msplinks page, that still won't result in a bidirectional
link...

-Dale
> --~--~---------~--~----~------------~-------~--~----~
> You received this message because you are subscribed to the Google Groups
> "Social Graph API" group.
> To post to this group, send email to social-g...@googlegroups.com
> To unsubscribe from this group, send email to
> social-graph-a...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/social-graph-api?hl=en
> -~----------~----~----~----~------~----~------~--~---

--
You received this message because you are subscribed to the Google Groups "Social Graph API" group.
To post to this group, send email to social-g...@googlegroups.com.
To unsubscribe from this group, send email to social-graph-a...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/social-graph-api?hl=en.

Brad Fitzpatrick

unread,
May 11, 2010, 1:55:19 PM5/11/10
to social-g...@googlegroups.com
SGAPI is somewhat understaffed at the moment, as I spend most of my limited time on Android performance lately.  So unlikely...  :-/

Dale Newfield

unread,
May 12, 2010, 11:42:11 AM5/12/10
to social-g...@googlegroups.com
LinkedIn does use rel="contact" for the three contacts it shows on your
profile page, but since there's no rel="me next" on the link to "See all
Connections" (and links from one page of connections to another), and no
way of getting the site to display a rel="me" link back to any external
pages, there appears to be no way to get SGAPI to leverage from those
carefully collected contacts. Just like myspace has an intermediate
page on outbound links, linkedin inserts an intermediate redirect:
http://www.linkedin.com/redirect?url=http%3A%2F%2Fnewfield%2Eorg%2Fdale&urlhash=fBF1

If they're trying to learn what outbound links are followed, they could
do <a href="external site"
onclick="window.location=redirect?url=external+site;return false">My
Website</a> so as to continue getting that info w/o breaking automated
link inference. Of course, since they don't include rel="me", they
clearly don't care about that.

Anyone here work at linkedin?

Is this project withering on the vine due to these walled gardens?

-Dale

On 5/11/10 1:55 PM, Brad Fitzpatrick wrote:
> SGAPI is somewhat understaffed at the moment, as I spend most of my
> limited time on Android performance lately. So unlikely... :-/
>
> On Tue, May 11, 2010 at 10:49 AM, Dale Newfield <da...@newfield.org
> <mailto:da...@newfield.org>> wrote:
>
> So, any chance that this nearly two year old request might get
> implemented, so that I can wind up with a bidirectional rel="me" to a
> myspace page, and thus be able to leverage the "friends" specified
> there?
>
> Specifically, on my myspace page http://www.myspace.com/dnewfield I
> added <a href="http://newfield.org/dale" rel="me">, but of course they
> replace that with <a
> href="http://www.msplinks.com/MDFodHRwOi8vbmV3ZmllbGQub3JnL2RhbGU="
> rel="me">, and while I could add a rel=me link from my real home page
> to this msplinks page, that still won't result in a bidirectional
> link...
>
> -Dale
>
> On Tue, Jul 8, 2008 at 2:48 AM, Brad Fitzpatrick
> <brad...@google.com <mailto:brad...@google.com>> wrote:
> > On Mon, Jul 7, 2008 at 11:36 PM, Martin Atkins
> <ma...@degeneration.co.uk <mailto:ma...@degeneration.co.uk>>
> <mailto:social-g...@googlegroups.com>
> > To unsubscribe from this group, send email to
> > social-graph-a...@googlegroups.com
> <mailto:social-graph-a...@googlegroups.com>
> > For more options, visit this group at
> > http://groups.google.com/group/social-graph-api?hl=en
> > -~----------~----~----~----~------~----~------~--~---
>
> --
> You received this message because you are subscribed to the Google
> Groups "Social Graph API" group.
> To post to this group, send email to
> social-g...@googlegroups.com
> <mailto:social-g...@googlegroups.com>.
> To unsubscribe from this group, send email to
> social-graph-a...@googlegroups.com
> <mailto:social-graph-api%2Bunsu...@googlegroups.com>.
Reply all
Reply to author
Forward
0 new messages