Amir Plivatsky

`build_clause()` maintains 2 totally different disjunct costs - one for the purpose of cut-off and one for computing the sentence metrics. See [table2](https://www.semanticscholar.org/paper/Predicate-Expression-Cost-Functions-to-Guide-Search-Bottaci/2ae3316f016f52c83e32e7b2ff7ef4abeb155c1b) and the article itself for discussion of...

I would like to push soon my patch to the SAT-parser for handling disjunct cost. Hence I would like to understand the rationale behind the apparent contradiction between the actual...

> I would like to push soon my patch to the SAT-parser for handling disjunct cost. So I will do it in a way that is compatible to how it...

Nothing to do on this issue for now, so I'm closing it.

By looking on the expressions generated by problematic db dictionaries, I realized that building disjuncts could be sped up if the expressions would get simplified. I specified 2 places in...

I'm confused regarding the needed fix. See "**EDIT**" in the original post above. My thought that it depends on the place the cost is distributed is not correct, since the...

In an old post above I said: > See table2 and the article itself for discussion of why max() is used. According to it, a disjunct cost is the maximum...

Just now I tested using `cost` instead of `maxcost` as a cost-cutoff. Since `maxcost

Here is the changed loop: ``` for (c4 = c2; c4 != NULL; c4 = c4->next) { //double maxcost = MAX(c3->maxcost,c4->maxcost); if (e->cost > ct->cost_cutoff) continue; c = pool_alloc(ct->Clause_pool); c->cost...

My change to line 177 is not optimal. However, it doesn't change the correctness as long as a too-low cost is used (this line is for efficiency only). It should...