From: "trans.  (T. Onoma)" <transami@...>
Date: 2004-11-01T01:10:20+09:00
Subject: Re: Flea vs RubyGarden

On Sunday 31 October 2004 10:28 am, Phlip wrote:
| Chad Fowler wrote:
| > Phlip wrote:
| > > Thanks - I just "got" it. To rescue my few pages, I must edit _all_
|
| their
|
| > > http: tags.
| > >
| > > When you change a page and save it, any http: tags bounce, not just the
|
| ones
|
| > > you edited.
| >
| > That's right.
|
| Okay. Imagine a spammer changes their URLs (in their input files!) to read
| HTTP:
|
| Then they attack.
|
| Any page with http: in it, they can't edit. They must erase the content, or
| upgrade its http:s to HTTP:s. The later is unlikely.
|
| However, my pages, which I just upgraded to HTTP:, are now _more_
| vulnerable to attack.
|
| >sigh<

You bring up an interesting insight. If a page contains a "mark" of 
un-save-ability, then it useless to spammers unless they can identify that 
mark and remove. We've already seen that something as simple as 'http://' can 
help do that. Of course eventually they could easily circumvent that. But we 
have an interesting solution in the making.

If we allowed a special mark for a page, such that the lock sequence would 
have to be _removed_ from the page before it will save. This would largely 
thwart spammers b/c they generally just _add_ to the page --moreover, if the 
lock sequence is redefinable to some degree it may help even more.

But how might it be done?

T.