I don’t know if ARBIL is still under development, but I’ll give it a try:
We have severe performance problems using ARBIL instead of the old IMDI Browser.
For example when opening the corpus IMDI / BAS / ALC
ARBIL starts reading to complete hierachy below the corpus node (the session nodes). It takes about 10 min memory gathering until ARBIL ‘hangs’, probably simply slowed down because of memory problems, and must be killed.
The old IMDI Browser seems to load only content that is actually opened (needed), so it works fine until you have opened to many sessions. Then it slows down, too.
Since ALC is only a medium sized corpus, this means that users cannot use ARBIL to brows larger BAS corpora.
What can we do about that?
Best,
Florian Schiel
Currently there is a known issue with the “installed version” under Microsoft Windows. Arbil can use more memory (and perform better) under Windows when you use the “webstart” way.
I tried importing “BAS ALC” and it did take some time (30 min in a virtual machine), but it worked (although there were only a lot of validation errors).
There are two separate issues here. One is being addressed with a lazy loading process, which means that larger corpuses can be viewed with less memory/time. While the other appears to be about the memory allocation in the windows installer version (a known issue).
Both of these separate issues have been addressed respectively, but have not made their way into the stable version as yet.