All right, thanks for the outpouring of support. I will explain briefly
some of what I had been trying to accomplish. I had been asked by Dave
Case and in response to requests on the reflector to get some tutorials
together for the force field applications I've been working on. I want
other Amber Devs and the community as a whole to have these things. I took
some time this past week to get one of them together and I'm nearly done
with the second. However, when I went to the website to add them, I had a
bit of trouble placing them and I assessed the current state of the
tutorials as follows:
- Single page format with all 49 abstracts is too long.
- It is hard for new contributions to gain visibility.
- It is hard for new users to navigate the tutorials we offer and find what
they want.
- The overall appearance of the site should change and one corner might be
a good start.
I am willing to work on the tutorials if people are willing to give
constructive feedback and allow me a chance to respond.
Dave
*- Single page format is too long.*
We may navigate with Ctrl+F and know exactly what keywords to search
for, but first year graduate students who need the tutorials the most do
not. They are mouse oriented, they read stuff, and they will come to the
site with a sense of what they want to do but not specifically what it's
called. Google "worst family feud contestants." When you read and search
for things like what you're trying to do, you get bogged down if there are
49 abstracts to parse through. The first five you might read in full, then
you'll skim, then you'll eventually take anything that sounds remotely like
what you want. The internal links at the top of the page help the problem
a bit, but they're not always descriptive of what's inside, and even then
you can go down to one section and have to scroll some more.
There are also logistical considerations. If everything is on one
page, that means that all 49 ToC images have to be transferred when the
page loads. People look at this site on their mobile devices, and we've
all had the experience of mollasses internet when at conferences. (This is
definitely a time I can think of people wanting to see an Amber
tutorial--after hearing a talk when they want to try out our new methods.)
After going through the page to the extent I have, I can see that one of
those images is really a large animation and others are bigger than the
monitor before the page resizes them. That's a lot of download!
Some of the image files should be resized, but others like Jason's
constant pH illustration are really well done and should be given more
space to tell their story. This is why I moved to unbind the image sizes,
but this will lengthen the page, which is another reason to break it into
sections.
We don't like code libraries to be more than 500 or 1000 lines
long--why web pages? And, even if we developers are comfortable with the
marathon page, most websites try to avoid that sort of format. I will,
however, limit the way new tabs are opened. While I think it would be nice
to spawn new tabs for specific tutorials, I was trying to find a balance in
my own tutorial and I went overboard. This is all easily fixed by changing
the CSS.
*- It is hard for new contributions to gain visibility.*
Let's say you make a new tutorial, and you roughly place it among one
of the categories. Where in the list do you place it, though? At the
front of the category? At the back? What if it's a new way to accomplish
something that an older method with a tutorial already addresses? (Could be
force fields, could be free energy estimates, enhanced sampling is this way
too.) The problem is that, without some unbiased way of presenting all of
the ideas at once, you'll want to be the first in the list to garner the
most attention, or perhaps the last in the list to look like the final
word. And no matter what you do, in the current format all you've got is
your abstract somewhere in the pile, with no changes at the top or
elsewhere on the site to direct people to your work. A ToC with only
titles, which presents everything in a handful of lines that can all be
viewed together, is the fairest way I can think to solving this problem.
The goal is to lift all boats.
If you think it is too much work to maintain two entries for each
tutorial, please consider that forty people over the course of ten to
fifteen years have compiled a list of 45 abstracts, and while there are
doubtless other tutorials that have been discarded, it only takes a minute
to copy the title and make a link to the abstract. I did it for all of
them yesterday afternoon. This is not a maintenance issue.
*- It is hard for new users to navigate the tutorials we offer and find
what they want.*
Already discussed; we need some means of quickly offering a list of
all tutorials that fits on one or two screens with minimal scrolling. If
we take a minute for each tutorial to maintain an actual ToC in addition to
the list of abstracts, that saves many users the hassle of deep searching
for a tutorial that may or may not exist.
*- The overall appearance of the site should change and one corner might be
a good start.*
The paper.gif background harkens to 1998 and our software is
definitely modern. Even in 256 browser-safe colors we can find
combinations that are more appealing and evoke "amber." But, more
importantly, the fonts need to change. The Sans-Serif headings contrasting
with the serif text is odd, but the italicized blue serif links do not fare
well on top of the textured grey background. This is a computational
chemistry package, and we often express ideas in which commas and tick
marks matter. They get lost against the textured background and they can
blend in with serif fonts, especially if italics are in play. People who
know the material may not see a problem, but they're also not the ones who
need to read the pages.
I don't tout the aesthetics of what I made: rather, I was going for
complete neutrality. We can dress it up later, but the intention was to
get away from the existing background and typeset in one of the parts of
the site where it matters the most, to allow experimentation and get
feedback.
Anyway, so that's my case. I don't actually feel that passionately about
it, but it seems the genie's out of the bottle, so if you truly feel that
you have something to say about Amber tutorials, I will take your
suggestions.
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Thu Sep 29 2016 - 16:00:02 PDT
