Methodology

Methodology

This page documents how the data shown on Conference Partner is sourced, what each ranking system means, and how often everything is refreshed. The goal is to be explicit about what we own (the dataset, the editorial workflow) versus what we reproduce (rankings published by external bodies, impact factor values published by publishers).

1. Conference and journal records

Each conference has a permanent canonical page at https://www.myhuiban.com/conference/{id} and each journal at https://www.myhuiban.com/journal/{id}. The set of fields shown per entity is:

  • Conferences: short name, full name, official website, submission deadline, notification date, conference date, location, edition number, CCF / ICORE / QUALIS tiers, and the full CFP text.
  • Journals: full name, optional short name, official website, publisher, ISSN, impact factor, and CCF tier.

Authoring is community-driven: any registered user can submit edits to the conference / journal page, with an audit line showing "Last updated by … in YYYY-MM-DD" on every detail page.

When a conference is renewed for a new year, its previous edition is snapshotted into a history record, so a venue accumulates a chronological series of editions over time rather than overwriting the past.

Each venue is additionally auto-tagged with research fields (14 computer-science areas such as AI & machine learning, security & privacy, systems & architecture) by matching a curated keyword dictionary against its full name. These tags power the field filters on list pages (e.g. /conferences/field/ai) and in the deadline calendar. They are heuristic labels we generate ourselves — not part of any external ranking system.

2. Ranking systems

2.1 CCF (China Computer Federation)

The CCF catalogue classifies computer-science conferences and journals into A, B, C, with A being the most selective. It is maintained and periodically revised by the China Computer Federation. Conference Partner reproduces the published CCF tier for each venue; we do not define the ranking. A value of x in our data means "not present in the current CCF list".

2.2 ICORE (Computing Research and Education Association of Australasia)

The ICORE conference ranking uses A*, A, B, C tiers, with A* the most selective. It is maintained by ICORE Inc. (Australia / New Zealand) and updated periodically. Internally we store A* as the string 0 for sortability; the display layer always renders it as A*.

2.3 QUALIS (Brazilian CAPES)

QUALIS is published by Brazil's CAPES agency (Ministry of Education) and uses tiers A1, A2, B1, B2, B3, B4, B5, C, with A1 the most selective. We reproduce the published tier per venue.

2.4 Relationship between the three

The three ranking systems are independent. A given conference can hold different ranks in different systems (for example, CCF-A and ICORE-B). We display whichever ranks are available for that venue. We do not compute or display a unified score.

3. Impact factor

Impact factor values shown on journal pages are reproduced from publishers' self-reported figures. They may lag the latest Journal Citation Reports (JCR) release. For authoritative current values, consult JCR directly.

4. Update cadence

  • CFP data: updated whenever a registered user submits an edit on a conference / journal page; see the audit line at the bottom of each "Call for Papers" card.
  • Ranking tier data: refreshed when the upstream body publishes a new list — typically every 1–3 years for CCF and ICORE.
  • Sitemap: <lastmod> values are derived from the underlying CFP / issue date per entity (not from the crawl date), so search engines and LLM crawlers see real freshness signals.
  • Application cache: public detail pages are cached for roughly 1 hour. Expect at most 1 h staleness from the underlying database.

5. Accounts, tracking and notifications

Anyone can browse the site without an account. A free account (with a one-time email verification) unlocks editing, tracking, comments and private messages. Tracking a conference or journal drives a set of opt-in email notifications, each carrying a one-click unsubscribe link and all manageable centrally from the account's email-preferences page:

  • Deadline reminders: sent 7, 3 and 1 days before a tracked conference's submission deadline, de-duplicated so each user gets at most one message per deadline per lead time.
  • New special issues: when a tracked journal publishes a new issue.
  • Conference updates: when a tracked conference's CFP changes.
  • Attendance reminders: 7 and 1 days before an event you marked as attending.
  • Weekly digest: a once-a-week summary of newly added conferences, journals and special issues.

Transactional and security emails (welcome, email verification, password reset, account changes) are always sent. Marketing and notification emails are sent only to verified, non-unsubscribed addresses.

6. Claiming for organizers and publishers

Conference organizers and journal publishers can claim their venue from its detail page. Each claim is reviewed by our team before it takes effect; once approved, the claimant gets a "My conferences" / "My journals" dashboard for the venues they manage. Claiming does not change how rankings or impact factors are displayed — those remain reproduced from the upstream sources described above.

7. Programmatic access

A read-only REST API is available with an OpenAPI 3 description at /api/openapi.json. List endpoints require a JWT obtained via POST /api/login (free account). The same read-only data is also exposed as a native MCP (Model Context Protocol) service at POST /mcp (Streamable HTTP, JSON-RPC 2.0) for AI agents: initialize and tools/list are open to anonymous clients, while tools/call requires either the same free JWT (subject to a daily quota) or per-request payment over the x402 protocol (HTTP 402). For each venue, free programmatic responses link back to its canonical Conference Partner detail page rather than an external website; paid detail calls additionally include the venue's official website URL. A site-level description for LLM agents is published at /llms.txt (with an extended snapshot including live-data samples at /llms-full.txt), and an RSS feed of deadlines in the next 30 days at /feed/upcoming-deadlines.xml. Conference and journal detail pages also answer requests with an Accept: text/markdown header (and no text/html) with a compact Markdown rendition built for agents — same data, no navigation chrome. An iCalendar subscription of upcoming submission deadlines is available at /calendar/deadlines.ics (subscribe by URL in Google Calendar, Outlook or Apple Calendar).

8. Out of scope

  • We do not host or distribute conference proceedings or PDF papers.
  • We do not own or define any of the three ranking systems.
  • We do not rank venues ourselves and do not compute any kind of composite "Conference Partner score".

9. Citation

When citing values from this site, use the canonical entity URL and attribute "Conference Partner (myhuiban.com)". If the claim is about a ranking value itself, cite the upstream ranking body instead — we are a reproducer, not the authority for those values.