mrgl-mrgl.com

A search engine that already knew you meant Warcraft

mrgl-mrgl.com was my attempt to make the World of Warcraft web easier to navigate. I built it in PHP to scrape, check, aggregate, and organize links about the game, bringing a scattered collection of community resources into a focused search experience.

The idea was rooted in a familiar player problem: knowing that useful information existed somewhere, but having to work out where to look. A game-specific search engine could start with that context already established. Its job was to help people reach the community’s material through a shared entry point.

I remember its beginnings around the late-2000s word-cloud era. The surviving public record gives the project a precise anchor: a working homepage captured on November 12, 2009. That distinction matters to this history—the recollection describes the period, while the archive establishes what was publicly visible on a particular day.

The interface that survives

The November 12 homepage preserves a compact search form with five categories: Web, Image, Video, Forums, and Items. It also includes Advanced Search and an email-to-a-friend feature. The page title identifies the service as Mrgl-Mrgl WoW Search, with a 2009 copyright notice.

Those categories express the product’s central design decision. A player looking for an item, a discussion, or a video should be able to narrow the kind of answer they wanted before following a result. Organizing the links was part of the search experience itself.

Word clouds are part of my recollection of the project, and cloud-tag styling survives in the source. The recovered November 12 homepage does not display a populated cloud, so this archive does not recreate one and present it as an original screenshot.

The PHP collection and search workflow

My work covered the PHP engine and the collection process behind it. Scraping and link checking were intended to make the aggregated results safer and more useful before presenting them to players. The surviving frontend establishes the search interface and PHP routing; the description of the backend comes from my own account of building it.

The architecture below reconstructs that workflow at a functional level. The original backend has not been recovered, so it does not claim a particular database, ranking formula, security mechanism, or crawling schedule. Link checking describes the system’s intent, rather than a guarantee that every historical result was safe.

When the beta escaped its intended audience

A November 13 update, captured the following day, records an unexpectedly public alpha/beta. The notice explains that a link intended for a small audience had spread, that bandwidth had run out, and that a server move was planned. It credits Geekologie and Geek.com with sending visitors.

Press coverage followed on November 15, 2009. Zach Yonzon’s beta report survives on Engadget.

A second notice dated November 16, captured November 18, describes further load problems and rebuilding work. It also announces a beta invitation for “Project Pond.” That name belongs in the record, although the recovered material does not establish its relationship to Snobold.

These notices preserve a useful part of the project’s story: a small community tool meeting the operational demands of public attention. They contain self-reported traffic figures, but no supporting analytics were recovered; those figures are not treated here as verified performance results.

Snobold: the successor that did not take off

snobold.com was my successor to mrgl-mrgl.com. It never took off, but it left more than a name behind. The February 3, 2011 archive preserves a page titled Snobold – WoW Search Engine and credits my character, Mayple, on Korgath.

The successor’s navigation points toward more specific player tasks: Websites, Image, Video, Wiki, Boss Tactics, WoW News, and AddOns. The frontend also contains live-search suggestion code and links to patch and community resources. These are surviving interface features, not proof that their backend services still work.

Snobold continued the same underlying idea: make a large community’s knowledge easier to find by organizing it around what its players need. Its limited outcome is part of that history too.

Preserving the work

Both projects are now historical. This page links to dated captures rather than sending visitors to the defunct domains. My résumé records MRGL as sold; the recovered material does not establish the buyer or transaction date.

What remains is a record of early web-tool development: PHP, collection and link checking, a search interface shaped around a game community, the pressures of an unexpectedly public beta, and a successor that pursued the idea further. It is an early example of a theme that still runs through my work—building tools that give useful information a clearer structure.

Historical note: this account combines my recollections with the linked archived pages and contemporary coverage. Capture dates establish when a page was saved, not necessarily when a project launched. Original HTML and asset provenance are retained with the portfolio’s research files.