Submissions

Two ways submissions arrive

Each assessment or question / sub-part is configured in Manage Assessments as either:

• On-ProfDesk, the student uploads their own file directly from their student portal.
• On-Paper, the student handed in physical paper; you bring it in yourself, either through AI Student Matching (bulk scan-and-match) or one at a time with Submit on Behalf inside ZenGrader.

Either way, once a file is attached to a student/question pair, it opens the same way in ZenGrader for grading.

The inline upload box on the Submissions tab appears only in the Best, Average and Worst candidate cells, because that page is for choosing work samples. To upload for any other student, open their assessment in ZenGrader and use Submit on Behalf; it then appears on the Submissions tab as a View button like any other submission.

When students can submit

An assessment accepts work between its available from date and whichever closing rule fires first: the deadline (if you asked it to close automatically there), or the available until date. A per-student extension pushes that student's closing moment past both. When a student cannot submit, the card names the reason, so a missed deadline reads differently from work that has not opened yet.

Closing an assessment never hides it. Students keep the instructions, their own past submissions and their grades; only new uploads stop. Reopen submissions puts it back on any plan, and it clears a date that is currently blocking, so a schedule you set months ago cannot leave you unable to reopen your own assessment.

Which file types an assessment accepts

By default a student may submit any file. ProfDesk stores it exactly as they submitted it and hands it back to you on download, whatever the format: a CAD drawing, a Pages document, a NumPy array, a PowerPoint deck, a zip of a whole project. Nothing needs adding to a list first.

The one exception is a program. A file an operating system runs or installs when you open it, such as .exe, .bat, .jar or .apk, is refused on every assessment, because you would be the one opening it. If a course genuinely collects those, ask students to submit a .zip instead. Source code is not affected: a .py, .sh or .js file is coursework and is accepted and shown as plain text.

Student Submission Settings on the assessment carries an Accepted file types card, directly under the submission method. Change opens a picker: tick the types this piece of work should take, or type any others into the Any other type field. PDF only is one click. Leave everything empty to go back to accepting any file. The rule appears on the student’s upload box, and a file outside it is refused with a message naming what is accepted.

Restricting the types is on every plan. It never affects work already submitted: tightening the rule after students have uploaded leaves their files, grades and annotations exactly as they are, and only applies to the next upload. An assessment split into parts sets this per part, and a part with nothing selected follows the assessment. Submit on Behalf is deliberately not bound by the rule, so you can always record a scan or a file a student handed you another way.

What ProfDesk can do with each type

Accepting a file is not the same as being able to work with it, so the picker marks each type with what it supports. AI means ProfDesk AI can read it, so AI grading works. S means the similarity checker can compare it. A type with no mark is graded by hand.

Three groups are worth knowing:

• PDFs, images, audio and video open in ProfDesk, so a recorded presentation plays without being downloaded. PDFs can be annotated on screen and are the only type the similarity checker compares.
• Text, data and source code open in ProfDesk as read-only plain text, so you can read a submission without downloading it. ProfDesk AI reads plain text, Markdown, HTML and CSV, but not code files.
• Everything else is stored and downloaded. Opening one in ProfDesk shows a download panel rather than a preview, because the app cannot render a CAD drawing or a slide deck. ProfDesk AI can still read a Word document.

The picker says which of these applies before you save, so an assessment restricted to code does not quietly turn out to be ungradeable by AI months later.

The Submissions tab and Sample Picker

Submissions & Work Samples shows per-assessment upload and view controls, plus the Sample Picker. The picker flags representative work at each performance level so the Course File Builder can include Best, Average, and Worst samples in the course record.

The Sample Picker groups performance into three tiers, each with a few label options so the wording matches how you talk about performance:

Tier	Label options	Meaning
Top	Best / Better / Good	A sample representing top performance on this assessment.
Middle	Average / Fair / Satisfactory	A sample representing median, typical performance.
Bottom	Worst / Poor / Weak	A sample representing bottom-tier performance.

ProfDesk suggests fallback candidates for each tier automatically based on scores, so you always have options. Uploading/flagging a file for any one candidate in a tier fulfills that tier's requirement, and the option is hidden for the other candidates in the same group, so you only need to pick one representative sample per tier, not one per candidate.

Where this shows up later

Flagged samples feed straight into the Course File Builder's performance sections, so your accreditation-ready course file has concrete evidence attached, not just scores.