The set syntax is older than the object notation.
The advantages of the set syntax are that you can use it in older Matlab releases, you have more flexibility in how to assign property names, and that you can use incomplete names and incorrect capitalization.
To show the flexibility, try running this:
clc
f=figure(1);
set(f,'nam','abc')
try
f.nam='foo';
catch ME
disp(ME.message)
end
try
f.name='bar';
catch ME
disp(ME.message)
end
However, this does come at the price of some performance: set seems to be about twice as slow, depending on the task. Below you find a tester that shows the performance for two simple tasks. The timings are a bit misleading, because graphics updates will be machine/OS depedent and will introduce more lag. This extra lag should be the same for either syntax.
clc,close all
handles=cell(1,100);
for n=1:size(handles,2)
handles{n}=figure('Name','abc','Position',[0.2 0.2 0.2 0.2]);
end
tic
for n=1:n
set(handles{n},'Name','foo')
set(handles{n},'Position',[0.3 0.3 0.2 0.2])
end
t_set=toc;
tic
for n=1:n
handles{n}.Name='bar';
handles{n}.Position=[0.1 0.1 0.2 0.2];
end
t_obj=toc;
close all
fprintf(['set syntax takes about %.2f ms per iteration,\n',...
'object notation takes about %.2f ms per iteration\n'],...
t_set*1000/n,t_obj*1000/n)
This returns this on my machine:
set syntax takes about 0.19 ms per iteration,
object notation takes about 0.10 ms per iteration