On Dec 5, 5:59 am, Michael Campbell <[email protected]>
wrote:
>  I think now with the new enhancements to
> cramming that I helped to propose some of that can be taken care of; it's
> like a mini-scheduler.

Are you referring to Anki's new "study sessions" such as "Review
Ahead?" In some ways I like this. For example, configuring the study
session to not update the card's schedules. Compared to v1.2, that's
really nice. 1.2 would update the schedule, sending due dates further
and further into the future, making "Review Early" not too useful for
"cramming." (I know it was never intended for that. But, its title led
everyone to reach for that as a way to study more frequently than
daily.).

But, I don't know yet how useful it really is. I disable schedule
update, and I see the same cards in the same order over and over
again. It's not really feeding back my scores, reprioritizing and
reordering cards for the next cram session. I can change the ordering
to "random," but then I'm not benefiting from SRS's prioritization of
cards (focusing my energy on cards that need more work).

To me, that seems like the persistent frustration I feel. There's
always a large gap between "perfect world" SRS scheduling and
pragmatic cramming. In 1.2, if you managed to discover the per-day
schedule switch and micro-interval assignment (and, Adam's "Forgetting
Index" plugin to retard the growth of "Young" card intervals) it was a
fairly seamless *continuum* from learning and longer-term retention
testing. If you wanted to review early, start 1.2 and review whatever
was due at that moment. If not enough was coming due, use the "Cheat"
plugin to reschedule a block of cards (the next 3 days) to be due
between now and 3 days.

The "learning mode" and "steps" seems ok. But, it doesn't seem to
dynamically adjust schedules like applying SRS (time, ease factor) to
micro intervals. It seems to create a hard either/or, not a natural
flow between learning versus mastery retention. And, that dividing
line between the two seems arbitrary (steps you pull from thin air
rather than a sliding scale of scheduling).

IMO, 1.2 was almost perfect in this regard. All it needed was to 1)
factor actual "FI" performance into the user's desired "FI" goal, and
2) make the rescheduling tool more of a "window" dragging function.
For example, specify the left and right boundaries of a timeframe and
drag the left boundary nearer to now, causing the affected cards to be
rescheduled *relative* to each other, maintaining their order. Perhaps
a center handle to nudge everything in the window more near-term.

(Rescheduling tools seem to define a window, but "sprinkle" the cards
within that window, disregarding the order/priority which has
developed between cards. For example, I might select a group of cards
and say "reschedule between today and tomorrow." A card due in 3 days
might "sprinkle" into today, and a card due today might "sprinkle"
into tomorrow. I think it should be more like condensing or expanding
the existing schedule into a new timeframe. Preserving order/spacing.
Factoring a set of cards into a new timeframe, not randomly seeding
them into new positions.).

To me, that would make "Review ahead" more useful for cramming. You
could drag the schedule back to near term and review ahead, gaining
the benefit of prior reviews reprioritizing/ordering cards without
them moving further into the future.




-- 
You received this message because you are subscribed to the Google Groups 
"mnemosyne-proj-users" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to