Dan
Sounds like you may be using EA in a manner similar to me. I'm guessing. I could be wrong. And probably am.
We use Component diagrams to model baseline and target application integration architectures. Component = application. Information Flow connector = interface.
We use a UML Profile to define tagged values for components and connectors, and we populate these fields with (more or less) structured information such as technology, business importance, business owner, etc. Enums where possible.
"Unstructured" textual information we place in the "Notes" compartment.
On the connector, we identify the type of business data flowing over the interface. We model the direction of flow with an arrowhead toward to "consuming" application.
The definition of what actually constitutes an "interface" - and how, in practice, one represents an "interface" in terms of level of granularity and directionality, etc - can turn out to be more complicated than the casual observer would initially think.
BTW - I'm not talking about SOA services here - just your traditional data-shipping, MOM and API style interfaces, which is where most organisations still are today... (Approach would still work, I think, I just haven't yet done this kind of modelling for a client that implements wholesale SOA services...)
Works a treat.
You can end up with quite large and complex diagrams in a large organisation, but the ability to search (control - F) and find things quickly is valuable, and the RTF document generation (flawed as it is) is useful.
The Relationship Matrix capability is also powerful.
We always create one single, overarching "everything including the kitchen sink" component diagram, which we let get as ugly as it gets, and don't worry too much about trying to make it presentation pretty.
This is called the "Big Hairy Ugly" diagram, and it exists for a purpose.
But then, we create separate segment / capability or domain diagrams. These are inherently less complex and can be made to look more visually comprehensible. These diagrams support "real architecture work".
I call the result:
"A machine for answering inevitable, ubiquitous, and recurrent architecture questions about applications and integration."
FWIW - I tried playing around with the "port" interface on components but found it too fiddly and no value added. You may feel differently.
The word "schematic" is relevant for these diagrams. However, it's possible that you may be using EA to actually create
electrical diagrams.

. If so, it's a pretty clever use of EA, I must say. And it will probably work, too...
In future, I intend to simplify things even more, by not using the Information Flow connector - just using a simple "Dependency" connector.
Now. Inserting an image.
You can insert a URL that links to a website, or a URL that links to an image on a website.
If an image - paste the full URL of the jpeg into your message, then highlight the entire URL, then click on the icon that looks like a small picture, which you will find above the text box that we use to type these messages (popup will say "insert image").
You will need to FIRST upload your image to your own website, or a public hosting website such as Mediafire, Picasa, Flickr, etc. and link to that URL...
Hope this helps ...
Jon