The portal (sardar) of the Yusuf-Khwaja Madrasa, known as Do Dar, in the shrine complex of Imam Reza in Mashhad, is one of only two surviving Timurid madrasas in the city and the only one whose portal decoration retains substantial elements of its original ninth-century-AH (fifteenth-century-CE) fabric. Unlike the neighboring Parizad Madrasa, whose portal inscriptions and tilework were entirely replaced in a Safavid-period repair, the Do Dar portal preserves an inscribed foundation text, a stone repair inscription from the reign of Shah Soleyman Safavi, and fragments of its original tilework—yet its façade is now bare plaster, its poetic inscriptions are largely lost, and a later gypsum muqarnas conceals the vault beneath the main inscription band. In the absence of any complete written or pictorial record of the portal's original appearance, restoring its decorative program raises two linked problems: first, how can the original decorative structure be reconstructed with scholarly defensibility when no direct documentation survives; and second, what can a comparative stylistic analysis of contemporaneous Timurid buildings responsibly establish, as opposed to merely conjecture?

This article addresses these questions through a comparative stylistic (typological) method, examining the Do Dar portal alongside a corpus of Timurid-era iwans in Khorasan and Transoxiana—including the Ghiyathiyya Madrasa at Khargird, the Masjid-i Shah in Mashhad, the shrine of Abu Bakr Taybadi, the Gowharshad Mosque, the (now-destroyed) Balasar Madrasa, and the Gur-i Amir and Tuman Aqa complexes in Samarkand—selected for their temporal proximity and shared function as inscribed entrance iwans. Two complementary bodies of evidence supplement this comparison: historical photographs held in the archives of the Cultural Heritage Organization and the Astan-e Quds Razavi Documentation Center, and the nineteenth-century account of the portal's inscriptions preserved in Mo'tamed al-Saltaneh's Matla' al-Shams.

The analysis identifies three distinct organizational types for the interior of Timurid iwans—a "mihrab-like" (scroll) type, in which a single continuous cartouche traces the arch's profile; a "girdled" (belt) type, in which a horizontal cornice inscription divides the iwan into upper and lower registers, leaving the vault itself always undecorated by muqarnas or stalactite work; and a "blind-arcade" type, confined mainly to secondary iwans within mosque courtyards. Comparison shows that the Do Dar portal's interior corresponds precisely to the girdled type, which implies that its present gypsum muqarnas—far from being original—is a later, undocumented addition inconsistent with Timurid decorative logic. This inference is independently corroborated by two lines of physical evidence: a small aperture accidentally left in the plaster revealing intact underlying tilework of a six-pointed star pattern, and the archaeological discovery, upon partial removal of the muqarnas, of brickwork framing a rectangular opening at the center of the tympanum—evidence of an original latticed tile window analogous to that surviving at the Masjid-i Shah. A comparable methodological pattern emerges in the article's treatment of the façade: comparison with the destroyed Balasar Madrasa, the closest analogue in both date and social rank, indicates that the Do Dar façade most plausibly followed a "niche-and-panel" organizational scheme rather than the scroll-cartouche scheme typical of royal foundations—while a discrete architectural anomaly, an asymmetric arched tympanum above the door with no parallel in authentic Timurid gateways, is shown to be a later alteration requiring redesign rather than restoration.

A parallel case study reconstructs the poetic inscriptions flanking the entrance, attributed to Sa'di, by cross-referencing surviving tile fragments, Mo'tamed al-Saltaneh's incomplete nineteenth-century transcription, and the metrical structure of the published divan—demonstrating that convergence of independent, individually insufficient sources of evidence can yield a defensible reconstruction where no single source suffices.

Taken together, these cases substantiate a broader methodological distinction between documented reconstruction, which claims to recover a specific historical state and whose confidence can be calibrated against comparative evidence, and new design within documented stylistic principles, which makes no such claim to historicity but derives its legitimacy from conformity to established period conventions. The article argues that conservation projects routinely conflate these two categories under the single label of "design," and that distinguishing them explicitly—as demonstrated here through paired examples from the portal's inscriptions, tile panels, and architectural details—should be a standard component of the scholarly documentation accompanying any restoration intervention on historic fabric.