I've never used mathematica, and therefore know very little of the syntax.
But I recently started using wolfram alpha for simple things (you've got to admit it's nice!) and started learning the syntax very quickly when I wanted to change the range of my plot and had to click on "mathematica syntax" .... I figured out in two seconds how to change every parameter of the plot that was there.
Now .. what scares me... is that when more and more people start using wolfram alpha before having access to a symbolic computation software [ think about the high school students that use wolfram alpha to check their homework answers ] ... when they're deciding whether or not to buy maple or mathematica, they'll chose the one with the smaller learning curve, since they know so much about the syntax already.
Using mathematica over maple might also have other advantages with compatibility with wolfram alpha .. like .. what if you get a result and want to save it in a worksheet ??
I see wolfram alpha becoming a mathematical search engine like google is a web search engine... and used A LOT in the near future.
Already I think more people use mathematica than maple .. and in the end one of them WILL be obsolete. I hope it's not maple because I've spent so long learning it!
But I'm wondering .. how about the alogirthms ??
I think my symbolic computation professor convinced me to stop using sage by saying the algorithms were 20 years out of date.
Wolfram isn't a researcher in the field of symb-comp, like Keith Geddes is. I would think that Maple has more up-to-date, and efficient algoirithms than mathematica .... but in all my time searching for the answer to this, have not found one! Do any experts have anything to say here ?? I'm not sure which algorithms are better, Wolfram Alpha definitely seems faster not even considering that it's over the internet.