Historical Browser Profile Source Baseline

Status: active research baseline; no named profile authorized Last updated: 2026-07-26 Scope: period WAP handset-browser interaction evidence for future Waves compatibility profiles

Purpose and Boundary

This baseline records recoverable primary historical evidence for future device/browser profiles. It does not define WAP requirements, alter the neutral Class C Reference runtime, or authorize a named profile. All sources here are interop-reference evidence under Waves Source Authority Policy; only the locked normative corpus may create or redefine RQ-* requirements.

The research covers viewport metrics, typography, focus and roller behavior, softkeys and task placement, Options/menu ordering, editors and forms, history, timers, documented failures, and browser limitations. Archived folders and historical planning snapshots are outside this audit.

Evidence and Confidence Rules

Evidence classes used below are:

P1: manufacturer/browser-vendor publication with internal publication identity and notices.
P2: contemporary operator or platform-owner guidance based on named devices and browser tests.
S1: third-party transcription, manual mirror, or contextual recollection without a recovered publisher artifact. This can locate or corroborate evidence, but cannot close a profile gate alone.

Claim confidence is separate from source class:

high: the recovered publisher artifact states the behavior directly, or two independent primary publications agree.
medium: publisher-authored content is available only through a mirror/transcription, describes a browser family rather than one handset, or leaves an implementation detail implicit.
low: only partial service documentation, a secondary account, or an unresolved conflict exists.

An archived or mirrored payload establishes research availability, not permission to republish it. The recovered PDFs and extracted text remain outside Git. Hashes, sizes, publication identities, source routes, and research conclusions may be recorded here. No access controls were bypassed.

Recovered Source Index

Retrieval and verification date for all rows is 2026-07-26.

ID	Publisher source	Retrieval and integrity	Class / provenance	Redistribution posture
`OW-2000`	Phone.com, WML Application Style Guide: Designing and Developing WML Applications, `WMLS-01-001`, August 2000	2003 archive capture of the original Openwave URL; 398,527 bytes; SHA-256 `de13b6c11fd4033813b8ed3144f12e6eb41fc73d12c9b76f3b60a230ac5cda11`; PDF 1.2, 78 pages	`P1`; high content identity, medium byte lineage because no publisher checksum survives	Copyright notice reserves rights and restricts reproduction, modification, and distribution; private research only
`OW-2001`	Openwave, Graphical Browser Application Style Guide, Mobile Browser WAP Edition 5.0, `MBWS-50-002`, August 2001	public university SDK mirror; 544,068 bytes; SHA-256 `d8a911ddb1a3de095fcdd7a387c2ad622153f47bd8f0f13a6db4cba3a3dda65b`; PDF 1.3, 86 pages	`P1`; high content identity, medium byte lineage	Restricted copyright notice; private research only
`NK-7110`	Nokia, Service Developer’s Guide for the Nokia 7110, `9359203`, issue 3, November 1999	public mirror; 242,486 bytes; SHA-256 `9eb8a3a4d97266f31bd3806d53d60ea8661be656d647382d365b07bdc5e2fec4`; PDF 1.2, 34 pages	`P1`; high content identity, medium byte lineage	Notice permits personal download/printing only; private research copy retained outside Git
`NK-2000`	Nokia, WAP Service Designer’s Guide to Nokia Handsets, 16 June 2000	public university mirror; 354,355 bytes; SHA-256 `7440173d99f7dbad2b08cc832d212bdeb53cf984fff87623e81a54462514241f`; PDF 1.3, 16 pages	`P1`; high content identity, medium byte lineage; corroborates `NK-7110`	Notice permits personal download/printing only; private research copy retained outside Git
`ER-R380`	Ericsson, R380s Design Guidelines for WAP Services, `LZT 108 3339`, first edition, November 1999	public manual mirror; 667,004 bytes; SHA-256 `904dd9511ecc5605dc2fbbb35fdb7fefd518af4ac23b19b2308f9540d49ef3b0`; PDF 1.3, 25 pages	`P1`; high content identity, medium byte lineage	All-rights-reserved notice and no redistribution grant; private research only
`ER-R320-WP`	Ericsson, R320s White Paper, `LZT 108 3714 R1A`, first edition, March 2000	full-text mirror; publisher PDF not recovered	`S1` transcription of an identified `P1` publication; medium	Mirror view only; no binary, checksum, or redistribution grant
`ER-WAPIDE`	Ericsson, WapIDE 3.1 User’s Guide, 9 April 2001	full-text mirror; publisher PDF not recovered	`S1` transcription of an identified `P1` publication; medium	Mirror view only; no binary, checksum, or redistribution grant
`MO-P7389`	Motorola, Timeport P7382i/P7389i Level III Service Manual	full-text mirror; publisher PDF not recovered	`S1` transcription of an identified `P1` service publication; low-to-medium for browser interaction	Mirror view only; no binary, checksum, or redistribution grant
`GENIE-2001`	Genie, Application Style Guide for Openwave, Nokia 7110/6210/6250, Mitsubishi Trium, release 1.0, February 2001	page-by-page mirror; original publication was linked from Openwave’s developer library; publisher PDF not recovered	`S1` mirror of an identified `P2` publication; medium corroboration only	No verified redistribution grant; do not promote mirror content

The Wayback CDX record for OW-2000 reports a 240,928-byte capture record while the replayed PDF is 398,527 bytes. The replayed file has a valid PDF structure, matching internal publication identity, period PDF metadata, and the original archived URL, but the size discrepancy prevents a high-confidence byte-lineage claim without a publisher checksum.

For each recovered PDF, verification included file-signature and metadata inspection, page count, SHA-256, text extraction, and rendered-page review. The rendered Openwave softkey template, Nokia Options sequence, and Ericsson R380 browser layout were visually checked against the extracted text. Encrypted Nokia PDFs were inspected within their granted print permission; their copy restrictions were not removed.

Interaction Evidence Matrix

H, M, and L are claim-confidence levels; -- means the interaction does not apply to that form factor or no useful evidence was recovered.

Candidate	Viewport / type	Focus / controls	Softkeys / menu	Editors / forms	History	Timers	Failures / limits
Nokia 7110	H	H	H	H	H	L	H
Phone.com/Openwave 4.x family	M	M	H logical / L physical	H	H	M	H
Openwave Mobile Browser 5.0 graphical	M	M	H logical / L physical	H	H	M	H
Ericsson R320s	H	M	M	M	M	L	M
Ericsson R380s	H	H touch	–	H	H	M	H
Motorola P7382i/P7389i	M	L	L	M	L	L	M

Profile Assessments

Nokia 7110: ready for profile planning

NK-7110 and NK-2000 agree on the profile-defining interaction model:

96 by 65 pixels overall, with a 96 by 44-pixel content graphics area and four data lines;
proportional 8-pixel normal and bold fonts, header truncation, word wrapping, and underlined links placed on their own lines;
inverse-video focus, line-by-line scrolling, roller movement, and roller press to select;
left Options and right Back during normal browsing;
fixed Options ordering: Home, Bookmarks, current Select/Edit action, card do actions in source order, conditional Use Number, Empty Cache, then Exit with confirmation;
number, text, and password editors, with the right softkey changing between Back and Clear and password characters becoming masked;
prev/Back history behavior and documented bookmark behavior;
concrete rendering limitations, including flattened tables, ignored emphasis, images constrained or truncated to the 96 by 44 area, and no image-map links.

This is sufficient to plan fixtures and profile-specific goldens. It is not authorization to begin implementation before the Class C, compatibility-registry, and frame/input gates. A 7110-specific timer cadence, firmware/version matrix, and reproducible failure corpus remain open and must be treated as explicit gaps rather than inferred from general WML behavior.

Phone.com/Openwave: browser-family behavior recovered, handset identity still open

OW-2000 is the missing primary browser-vendor guide and provides strong logical behavior:

a typical four-line display, 12 to 15 characters per line, with deployed phones ranging from two to eight lines and using mostly variable-width fonts;
two programmable actions: the primary accept action and secondary options access, plus a fixed Back/Back-Clear control;
overflow actions placed in a numbered options menu, short labels, compact menus, and a preference for the common action on the primary binding;
complete guidance for text, numeric, password, and formatted fields, validation before acceptance, selection lists, retained values, and clearing entry state;
default Back/history and cache behavior, including activity/delete patterns that deliberately remove intermediate cards from history;
timer/image-loading races, unsupported-character substitution, no horizontal image scrolling, and significant OEM variation in display and control characteristics.

The guide’s display template says accept is on the left and options on the right, while adjacent prose describes the primary key as the right softkey. OW-2001 contains a comparable left/right inconsistency. Both guides also say handset manufacturers can vary browser characteristics. Therefore the logical primary/secondary precedence is high-confidence evidence, but a physical left/right mapping is not safe at browser-family level. A future profile must name one shipped handset and browser release, then pair the Openwave guide with that handset’s manufacturer user/developer guide.

OW-2001 is valuable successor evidence for the 5.0 graphical browser: it documents three to eight display lines, mostly variable-width text, link styling, pop-up menus, buttons, radio controls, and the rule that an activated editor or pop-up consumes the programmable-softkey surface. It must not be silently projected backward onto Phone.com 4.x or onto a handset whose OEM controls differ.

Ericsson: R320 evidence is promising; R380 is a separate touch profile

The R320 white paper provides a 101 by 65-pixel display, five full-screen rows including the header, four selection/input rows including the header, proportional font capacity, inverse link focus, an image focus frame, a 4 KB cache, and up to 25 bookmarks. The WapIDE guide adds short YES to follow or confirm, held YES for Options, short NO for Back/history, vertical scrolling, horizontal movement among table cells or multiple links, and a mixed menu of card actions plus fixed browser actions.

These are publisher-authored sources, but only third-party text views were recoverable. The exact R320 design guide, byte-verified developer PDFs, editor details, timer behavior, precise fixed-menu ordering, and firmware-specific failures remain open. The consumer manual also distinguishes Suspend in the browser menu from Resume in the phone’s WAP Services menu; they must not be presented as one fixed Options sequence.

ER-R380 is a strong primary source for a different device class: a 360 by 120 touchscreen with a 310 by 100 browser area, breadcrumb-like title/history display, toolbar navigation, proportional 9/10/14-pixel fonts, touch focus, list and button controls, and on-screen keyboard or handwriting entry. It should inform a later R380 touch profile, not fill R320 keypad/softkey gaps.

Motorola: service evidence is not enough for a named profile

MO-P7389 establishes a 96 by 54 display, WAP 1.1 support, simplified alphabetic entry, partial display of oversized bitmaps, browser entry through Quick Access/menu paths, and pause/resume around incoming calls. It does not establish focus traversal, exact softkey precedence, Options ordering, history semantics, timers, or a complete editor model. A manufacturer browser/application developer guide and a byte-verified consumer manual are still required before planning a Motorola profile.

Conflicts and Non-Inferences

Openwave’s logical primary/secondary action model is supported; its physical left/right mapping is unresolved and may be OEM-specific.
Openwave 5.0 graphical-browser behavior is successor/delta evidence, not proof of 4.x behavior.
Nokia 7110 and Nokia 6210/6250 corroboration does not make handset metrics interchangeable; the comparative Nokia guide lists different display heights.
Ericsson R320s keypad behavior and Ericsson R380s touchscreen behavior are separate profiles.
Suspend and Resume are different R320 state transitions in different menus, not adjacent fixed Options items.
Service-manual facts and third-party recollections are candidate-fixture context, not a basis for profile requirements.
None of these sources may redefine the neutral Class C Reference, WML task semantics, or a normative RQ-* item.

Readiness Decision and Future Work

The Nokia 7110 is the first named profile ready to plan because two concordant Nokia publications cover every requested interaction domain except device-specific timer cadence and a reproducible failure/version matrix. Planning must preserve those gaps and the existing runtime gates.

Openwave is next only as an evidence-lock task, not an implementation profile: select a representative shipping handset/browser version, recover its manufacturer interaction guide, resolve physical controls and editor modes, and establish device-specific viewport, timer, and failure fixtures. The browser-family guides alone do not justify a generic Openwave compatibility profile.

Ericsson R320 and Motorola remain source-recovery candidates. R380 has adequate design-guide evidence but represents a materially different touch device and should follow, not substitute for, the keypad profiles prioritized by the current product direction.

Unresolved Primary-Source Queue

A handset-specific Openwave/UP.Browser 4.x user or developer guide that names browser version, physical softkeys, viewport, editors, and known failures.
The publisher PDF for the Ericsson R320 design/developer guide, plus byte-verified R320 user and WapIDE manuals.
A Motorola application/browser developer guide for a representative WAP handset, plus a byte-verified consumer manual.
Device/firmware release notes or test reports for timer cadence, cache faults, network failure presentation, and known browser defects across the selected profiles.
Publisher checksums or authenticated archives for all mirrored PDFs, and explicit redistribution permission before any binary or extracted derivative is proposed for Git.