Author: Ratko V Tomic
Date: 09:03:07 12/01/00
Go up one level in this thread
>The calibration test is rather long (4 seconds) the first time you run the
>program. In the next sessions, it will be shorter (0.5 to 1 second), unless you
>change something in your hardware configuration (change in processor speed for
>example).
The loading lasts about 4.5-4.9 seconds every time (normally programs load on
this machine in under 1 second). The main waiting seems to occur after
LoadLibrary of engineps.dll (and before the next library is loaded) then there
are calls to create thread and Sleep(100ms) from the main CP5 thread. There is
also, a bit erlier, an exec (via CreateProcess) of regedit program with command
line to change something in registry (registry API's would do the job much
quicker).
>>The first bug above with overlapped windows is probably due to improper >>handling of window repaint when user has set his Explorer shell to
>>"Show window content while dragging" (have you tested window switching
>>under such shell setting?).
>
>I don't think it is, because I had the same problem under W95.
I am running it under Win98/SE. Was that problem ever fixed? The window repaints
in CP5 seem overall flaky, leaving bits of earlier windows or missing peaces of
windows, which clear up only after minimizing then restoring the top level CP5
window. I had similar repaint problems with another non-chess program (which was
in beta) and these were due to my shell being set to "show the window content
while dragging." In that case the programmer never saw the problem on his
machine since he had that Windows shell option turned off.
>>check the website "Software Hall of Shame" at:
>>
>> http://www.iarchitect.com/index.htm
>>
>>which has many commercial product examples of bad UI (including Microsoft's
>>products, Notepad and Explorer shell especially, and their
>>conventions mimicked by others). I think the future versions of CP
>> would benefit from learning a lesson from some of those examples.
>> However incredible it may sound to some, Microsoft doesn't always
>> know best (or means always the best, at least not for the end-user).
>
>They have the best record ever for the number of bugs in released
>products. From this point of view they have all my admiration.
The site above isn't about bugs but about usability (problems) of various
programs, how well the operation has been thought out and how much user's
conveninece and time mattered to the UI designer. For example, take a look at
Chessbase UI switching between database and play board - a single click, without
even having to move the mouse takes you back and forth. Now try the same with
CP5 UI -- switch back & forth few times, making sure all the play screen windows
come on top when dbase is to go away,... see how long it takes and how much
precision mouse moves and clicks it requires.
It is very obvious that the Chessbase designers carefully went over the
operation and put themselves into the user's place, looking for and smoothing
away all the wasteful (of time and effort) actions. For another example, the CB
main controls are at the bottom, since user will play from the bottom of the
board, so the mouse cursor will be closer (on average) to the controls. I have
impression that CP5 programmer was happy enough to struggle his way through the
mfc & ole, that any thought of such finesse as user's time and efforts for
innumerable high precision motoric actions (to hunt around and click on various
4mm squares, for just about anything you wish to do other than making moves)
never crossed his mind. And the same huge gap in attitude & attention toward the
end-user permeates every aspect of the two respective UIs.
The suggestion to the cp5 developers to check the site above was therefore not
meant to help them debug the cp5, but to help them see that there is a whole
another universe out there (beyond the "look ma, I can do mfc and ole") they
apparently aren't even aware to exist -- how to see things from the user's angle
(a user who has many other things to deal with, whose life and time do not
revolve around the particular product, as those of the developer, naturally,
do), to see very common oversights and types of clumsiness in various Windows
products, including their own. Another excellent related site, on the UI
usability is:
http://www.useit.com/alertbox/
which, although dealing mainly with Web design usability, also has good links to
general usability materials. These little things (users) do matter.
BTW, I did manage to squeeze enough time to play several games against the GT.
It is quite an enjoyable opponent. It is funny to see how from the get-go it
starts lining up toward my king and then suddenly, out of a blue, it throws away
a piece to win a game (out of about dozen games I won only 3, no draws; GT did
continue playing on well beyond the point at which Hiarcs or Fritz would have
resigned; there may be some setting for this, didn't look, didn't really need
such feature that much, anyway). GT does sometimes overreach with sacrifices,
which ends up costing him a game. I didn't have time yet for CT13 or RC3 (maybe
during Christmass holiday I'll try them). Although I haven't played against GT
under tournament tempo (I played only 5min +10 and +15 sec/move), from the few
games it seems that Hiarcs still has a better positional sense in the
non-king-attack positions, e.g. with GT I never got a sense of being
positionally squeezed with no good move left to play while the opponent can keep
improving his position (as it happens with Hiarcs). On the other hand, I never
had as many pieces sacrificed (successfully) against me in as few games as with
GT, before.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.