# org-mode HTML export - difference in behavior with org-html and org-pandoc

I have an example org-mode file with the following contents:

- only text\\
only text\\
only text
- image... (works with org-html but not pandoc-html because of no brackets around "file:" statement)
#+ATTR_HTML: :width 15%
file:emacsimage.png\\
and text
- image... (works with pandoc-html but does not respect HTML attributes)\\
#+ATTR_HTML: :width 15%
[[file:emacsimage.png]]\\
and text
- image... (works with pandoc-html but does not keep image or following text as part of this list item)

#+ATTR_HTML: :width 15%
[[file:emacsimage.png]]

and text


If I export using org-html-export-to-html, I get the following HTML file:

If I export using org-pandoc-export-to-html (from the ox-pandoc package), I get the following HTML file:

The main point is that for the #+ATTR_HTML tag to be recognized by pandoc (not an ox-pandoc issue), two conditions have to be met: 1) blank lines must be placed before and after the #+ATTR_HTML and figure specifications, and 2) #+ATTR_HTML must start at column 1 regardless of whether it is in a list item or not. This is true with other tags; e.g. #+BEGIN_QUOTE/#+END_QUOTE. This is not the case with the regular org-mode html export function.

Two questions:

1. Is it expected for the two to have different behavior? I thought there was only one org-mode specification (in contrast to markdown) as suggsted here (the "Org-Mode Is Standardized" argument, which I guess is not true?).
2. Is it possible to get the pandoc to behave like org-html with respect to these lists? Being able to place "and text" after the image in the second list item is more desirable as it belongs grouped with the preceding content. Pandoc's requirement for blank lines to be placed before and after the image insertion syntax prevents the image and anything that follows to be associated with the same list item. This also applies to other elements to be included in lists like #+BEGIN_QUOTE/#+END_QUOTE and so on.