About daftar

The problem it exists for

When someone asks what Islam actually says about something, they almost never get a sourced answer. They get confident assertion, a forum post, or silence. Meanwhile a great many practising Muslims would benefit from understanding how their fiqh works and why a ruling exists, and have no realistic path to that outside a traditional hawza.

Clarification — tabyeen — is not optional in this tradition. IlmFlow is a practical expression of it: the work of tabyeen made consistent and available, at two in the morning, to someone who has nobody to ask.

What this is not

These are design constraints written into the material itself, not disclaimers added afterwards.

• It does not issue fatwa or personal legal rulings.

  For any binding ruling it directs you to your own marjaʿ. Ijtihad is a spiritual and authoritative act, and it cannot be delegated to a machine.

• It does not claim to replicate any particular marjaʿ’s reasoning.

  It can set out a valid source-based pathway to a ruling, while saying plainly that it cannot know the scholar’s actual thought process.

• It does not replace hawza study.

  Pedagogy in this tradition is relational and developmental. Years under a murshid are not reproducible by asking questions of a system.

• It does not engage in political advocacy.

  It explains the scholarly basis of a position. It does not use religious authority as rhetorical leverage.

• It does not resolve genuine ikhtilaf by declaring a winner.

  Where qualified scholars differ, it presents the positions and the sources they rest on, and says which framework it is reasoning within.

• It does not handle irshad, duʿa, or the inner dimensions of worship.

  The spiritual dimension of the faith is not reducible to information retrieval.

What one record is

One question, one answer, and the sources the answer rests on — with the person who wrote it and the person who checked it both named. Not a passage copied out of a book: a worked teaching example, of the kind a teacher would give a student who asked.

Why a second person always checks

This is the part that matters most, and it is easy to underestimate. If an answer cites a page and nobody opened the book, what the model learns is not a wrong page number. What it learns is that producing a confident-looking citation is what an answer does — and it then applies that habit to every book it has never seen.

The failure is invisible afterwards: every output looks scholarly, and only someone checking the reference ever finds out. So a record with an unverified citation is held back rather than flagged, and forty checked pairs are worth more than four hundred unchecked ones.

What we cannot do ourselves

The technical work is ours. The scholarship is not — and none of it can be substituted with more engineering. It needs people who are hawza-trained: to write the first examples that set the register everything after them imitates, to review what follows, to say which texts and which editions belong, to build the glossary of technical terms in usul, fiqh and kalam across Arabic and Farsi, and to supply the loaded questions they have actually been asked.