Yahoo Mail is obsessed with blood

[steuard]

I sent Kim email today complaining about my dry skin and lack of lotion at work. In it, I wrote, 'I have this fear that my knuckles are going to progress from "dry" to "dripping blood" by dinnertime.' When Kim wrote back from her Yahoo Mail account, I happened to look at the raw source code of the message (different story), and where she had quoted my message I saw something odd.

The HTML-formatted copy of my quoted message included the following excerpt:

"...progress from "dry" to "<span class="yshortcuts" id="lw_1265039033_0">dripping<br>blood</span>" by dinnertime."

Yahoo has for some reason decided to insert some special HTML tags around my words "dripping blood", though neither Kim nor I observe any difference in how the text appears. There's also a lengthy block of associated JavaScript that I've included excerpts from below:

YAHOO.Shortcuts.hasSensitiveText = true;
YAHOO.Shortcuts.sensitivityType = ["sensitive_news_terms"];
YAHOO.Shortcuts.document_author = "<my email address snipped>";
YAHOO.Shortcuts.IBCategory = {"category":"travel", "score":"-0.694692"};

YAHOO.Shortcuts.annotationSet = { "lw_1265039033_0": {
"text": "dripping blood",
"weight": 0.51955, "relScore": 10.9459,
"type": ["shortcuts:/concept"],
"category": ["CONCEPT"],
"showOnClick": [],
"context": "going to progress from dry to dripping blood by dinnertime",
"metaData": {"visible": "true"} } };

I conclude the following:

• Yahoo has a long list of terms that make it uneasy.
• Yahoo assigns various scores and categories to each one (and tracks who said it).
• Yahoo is so proud of these assessments that it silently adds them into all outgoing mail just in case the recipient's email program wants to use them.
• Yahoo believes that "dripping blood" is always "news".
• Yahoo also associates "dripping blood" with "category: travel".
• Yahoo's programmers take some seriously messed-up vacations.
• Someone out there sees special formatting every time "dripping blood" is mentioned.
• You could probably do amusing things to them with the right "compatible" scripts.

Comments

[steuard.livejournal.com]

Ah, those terms do give some more information. I'd originally tried searching for the most "interesting" keywords (to me), like "hasSensitiveText", but that mostly turned up web pages with HTML emails regurgitated verbatim in the page text. (It looks like there are some blog-spam bots or systems that produce copious but absolutely useless output with this sort of markup.)

Searching for "Yahoo Shortcuts" gives some information, but I haven't spotted anything yet that addressed this "sensitive text" category. Most of the system seems targeted at providing search or shopping results.