[developers] Re: Macintosh version of LKB now publicly available
J.A.Carroll at sussex.ac.uk
Fri Jun 17 00:19:30 CEST 2005
On 16 Jun 2005, at 19:45, Woodley Packard wrote:
> I can confirm similar behavior. An additional clue: running "shrink"
> or "expand" on a node inside an AVM also takes a surprisingly long
> time on big ones, and seems to be correlated to the size of the AVM.
> I also noticed that occasionally dragging the scrollbars seem
> ignored... I parsed "the dog chased the cat" (under the ERG), opened
> up the top level AVM, scrolled all the way to the bottom, and then
> tried to scroll all the way to the top and found I was still looking
> at the bottom of the AVM (although the scrollbar indicated the top).
> Nodes were not clickable in this situation, though. If I shrink the
> window and make it bigger while in this strange state, the exposed
> portions get updated with the correct data, leaving orphaned data from
> the bottom of the AVM still sitting around in the other parts of the
I hadn't noticed that, but maybe I have a different setting for shrunk
((ARGS SIGN) (HD-DTR HEADED_PHRASE) (NH-DTR BASIC_BINARY_HEADED_PHRASE)
This is with the very latest ERG "LinGO (30-Apr-05; SP2)". With this
shrunk types setting, feature structures are quite manageable in size.
In fact, with *no* shrunk types, the feature structure window for the
parse of "the dog chased the cat" still behaves itself -- apart from
horizontal scrolling not quite reaching the most embedded part of the
I do get bad behaviour for longer sentences, such as "the dog chased
the cat that I saw in the garden", where it takes around 10 seconds for
the window to be displayed, and it comes without a vertical scroll bar.
But I don't get any strange artifacts in the window after resizing.
I was thinking that perhaps Emily was seeing her LKB session getting
bogged down because she didn't have enough memory in her machine? I
have 1GB in my powerbook G4 667HMz and it all seems to hum along quite
More information about the developers