An about page, and the attribution it owes

src/MUI.Crawl/ProbeOptions.cs

@@ -123,9 +123,26 @@ public sealed record ProbeOptions
     /// What the crawler calls itself over TTYPE/MTTS and MNES <c>CLIENT_NAME</c> (spec §11).
     /// </summary>
     /// <remarks>
+    /// <para>
     /// An admin reading their logs must be able to find out who we are and how to opt out, so this
     /// is a politeness obligation rather than a cosmetic string. It carries a URL for the same
     /// reason.
+    /// </para>
+    /// <para>
+    /// <b>It does not reach the wire yet, and neither does <see cref="InfoUrl"/>.</b>
+    /// TelnetNegotiationCore's <c>TerminalTypeProtocol</c> hardcodes a client's terminal types to
+    /// <c>TNC</c>, <c>XTERM</c>, <c>MTTS 3853</c> in a private field with no setter, and its
+    /// <c>NewEnvironProtocol</c> answers a server's NEW-ENVIRON request with the crawler host's own
+    /// <c>USER</c> and a fixed <c>LANG</c> — so what an admin actually sees is the library's default
+    /// and a local account name, not us. The library is first-party: the fix is a PR there making
+    /// both settable, never a reflection hack or a hand-rolled plugin here.
+    /// </para>
+    /// <para>
+    /// Until then nothing may claim otherwise. <c>/about</c> reads this field and says plainly that
+    /// the crawler is <em>configured</em> to call itself this and does not manage to, because
+    /// "the crawler identifies itself" is a claim about our own behaviour and that one would be
+    /// false in exactly the way <c>ContactedMaintainer</c>'s default was.
+    /// </para>
     /// </remarks>
     public IReadOnlyList<string> TerminalTypes { get; init; } =
         ["MUINDEX-CRAWLER", "MUINDEX", "MTTS 9"];
@@ -159,5 +176,11 @@ public sealed record ProbeOptions
     public const byte MsspOption = 70;
 
     /// <summary>Where an admin can read what we do and ask us to stop.</summary>
+    /// <remarks>
+    /// A placeholder domain, because the domain is an open question (spec §15.1) and inventing one
+    /// here would settle it by accident. It is also not yet sent to anybody — see
+    /// <see cref="TerminalTypes"/> — so a deployment that leaves this alone is publishing an address
+    /// that answers nobody. <c>/about</c> compares against this default and says so when it matches.
+    /// </remarks>
     public string InfoUrl { get; init; } = "https://muindex.example/crawler";
 }

src/MUI.Web/Components/Layout/MainLayout.razor

@@ -24,6 +24,7 @@
     <nav aria-label="Catalogues">
         <a href="/games">games</a>
         <a href="/archive">archive</a>
+        <a href="/about">about</a>
     </nav>
     <span class="kicker">every game here was checked by a machine, and every fact says when</span>
 </header>

src/MUI.Web/Components/Pages/About.razor

@@ -0,0 +1,123 @@
+@page "/about"
+@using Microsoft.Extensions.Options
+@using MUI.Crawl
+@using MUI.Web.Api
+@inject IOptions<DatasetLicenceOptions> Dataset
+@inject IServiceProvider Services
+
+@*
+    The page this project owes the sites it read and the servers it dials.
+
+    Spec §7.6 and §11 both require it, and neither of them calls it a feature: crediting the
+    directories a backfill read, publishing who is knocking and how to make them stop, and stating
+    the limits of what a measurement here proves are obligations already incurred by crawling.
+
+    It is prose and a short list, and it is deliberately the plainest page on the site — the whole
+    argument this project makes is that a fact stated with its provenance beats a fact presented
+    well, and the page making that argument would be a poor place to contradict it.
+*@
+
+@if (Plain)
+{
+    <pre class="mono plain">@PlainText.RenderAbout(Page)</pre>
+}
+else
+{
+    <PageTitle>About — mu*index</PageTitle>
+
+    <h1>About mu*index</h1>
+    <p class="summary">@Page.Lede</p>
+
+    @foreach (var section in Page.Sections)
+    {
+        <section class="about" aria-labelledby="about-@section.Id">
+            <h2 id="about-@section.Id">@section.Heading</h2>
+
+            @foreach (var point in section.Points)
+            {
+                <p class="note"><strong>@point.Lead</strong> @point.Body</p>
+            }
+
+            @if (section.Identity is { } identity)
+            {
+                <p class="note">@identity.Wording</p>
+
+                <dl class="identity mono">
+                    <dt class="kicker">crawler</dt>
+                    <dd>@identity.Name</dd>
+                    <dt class="kicker">contact</dt>
+                    <dd>
+                        <a href="@identity.InfoUrl">@identity.InfoUrl</a>
+                        @if (!identity.ContactConfigured)
+                        {
+                            @* Amber is the site's word for declared-but-unverified, and an address
+                               nobody configured is exactly that: a value with nothing behind it. *@
+                            <span class="warn"> — placeholder; this deployment set no contact address</span>
+                        }
+                    </dd>
+                </dl>
+            }
+
+            @if (section.Sources.Count > 0)
+            {
+                <ul class="sources">
+                    @foreach (var source in section.Sources)
+                    {
+                        @* Name and state on one line, the note under both. The listing's two-column
+                           row is wrong here: a note is a paragraph rather than a facts column, and
+                           reserving a column for a short badge leaves the prose reading in a gutter. *@
+                        <li>
+                            <p class="source-head">
+                                <a class="name" href="@source.Url" rel="external">@source.Name</a>
+                                <span class="badge">@source.StatusWording</span>
+                            </p>
+                            <p class="dim">@source.Note</p>
+                        </li>
+                    }
+                </ul>
+            }
+
+            @if (section.Licence is { } licence)
+            {
+                <dl class="identity mono">
+                    <dt class="kicker">code</dt>
+                    <dd>@licence.CodeLicence</dd>
+                    <dt class="kicker">data, as this deployment serves it</dt>
+                    <dd>
+                        @if (licence.DataLicenceUrl is { } url)
+                        {
+                            <a href="@url">@licence.DataLicenceName</a>
+                        }
+                        else
+                        {
+                            @licence.DataLicenceName
+                        }
+                    </dd>
+                    <dt class="kicker">credit as</dt>
+                    <dd>@licence.Attribution</dd>
+                </dl>
+
+                <p class="note dim">@licence.Notice</p>
+            }
+        </section>
+    }
+
+    <p class="faint"><a href="?plain=1">read this page as plain text</a></p>
+}
+
+@code {
+    [SupplyParameterFromQuery(Name = "plain")] private string? PlainFlag { get; set; }
+
+    private bool Plain => Truthy.Is(PlainFlag);
+
+    /*
+        The crawler's own options when the crawl loop is hosted in this process, and the defaults
+        when it is not. Asking the container rather than constructing one means the identity this
+        page publishes is the identity the probe was built from — a page that said MUINDEX-CRAWLER
+        while a deployment had configured something else would be describing the source tree instead
+        of the crawler, which is the failure mode of every claim on this site that nobody pinned.
+    */
+    private AboutPage Page => AboutPage.Build(
+        Services.GetService(typeof(ProbeOptions)) as ProbeOptions ?? new ProbeOptions(),
+        Dataset.Value);
+}

src/MUI.Web/Components/PlainText.cs

@@ -350,6 +350,79 @@ public static string RenderArchive(IReadOnlyList<ArchiveEntry> entries, string?
         return b.ToString();
     }
 
+    /// <summary>
+    /// The about page. Prose, so the only thing the graphical version adds is the shape of it.
+    /// </summary>
+    /// <remarks>
+    /// The attribution list is the part that has to survive here above all: it is what this project
+    /// owes the directories it read, and an acknowledgement a text browser cannot render is an
+    /// acknowledgement made to the layout rather than to anybody.
+    /// </remarks>
+    public static string RenderAbout(AboutPage page)
+    {
+        var b = new StringBuilder();
+
+        b.AppendLine("ABOUT MU*INDEX");
+        b.AppendLine();
+        Wrap(b, page.Lede);
+
+        foreach (var section in page.Sections)
+        {
+            Heading(b, section.Heading);
+
+            foreach (var point in section.Points)
+            {
+                b.AppendLine();
+                Wrap(b, point.Sentence, "  ");
+            }
+
+            if (section.Identity is { } identity)
+            {
+                b.AppendLine();
+                Wrap(b, identity.Wording, "  ");
+                b.AppendLine();
+                Wrap(b, $"Crawler: {identity.Name}", "  ");
+                Wrap(b, $"Contact: {identity.InfoUrl}", "  ");
+
+                if (!identity.ContactConfigured)
+                {
+                    Wrap(b, "This deployment has not set a contact address, so the one above is "
+                        + "the built-in placeholder and answers nobody.", "  ");
+                }
+            }
+
+            foreach (var source in section.Sources)
+            {
+                b.AppendLine();
+                b.AppendLine($"  {source.Name} — {source.StatusWording}");
+                b.AppendLine($"  {source.Url}");
+                Wrap(b, source.Note, "    ");
+            }
+
+            if (section.Licence is { } licence)
+            {
+                b.AppendLine();
+                // Every one of these goes through the wrapper rather than being laid out in columns:
+                // a licence name and an attribution are both configuration, and a deployment that
+                // sets a long one must not push a line off the side of a text browser.
+                Wrap(b, $"Code: {licence.CodeLicence}", "  ");
+                Wrap(b, $"Data: {licence.DataLicenceName}", "  ");
+
+                if (licence.DataLicenceUrl is { } url)
+                {
+                    Wrap(b, url, "  ");
+                }
+
+                Wrap(b, "(what this deployment serves. The project's own answer is still open.)", "  ");
+                Wrap(b, $"Credit as: {licence.Attribution}", "  ");
+                b.AppendLine();
+                Wrap(b, licence.Notice, "  ");
+            }
+        }
+
+        return b.ToString();
+    }
+
     private static void Heading(StringBuilder b, string title)
     {
         b.AppendLine();

src/MUI.Web/wwwroot/app.css

@@ -517,3 +517,21 @@ fieldset.facets .check { display: inline-flex; gap: 6px; align-items: center; fo
     border-bottom: 1px solid var(--amber);
     color: var(--text);
 }
+
+/* ── about ─────────────────────────────────────────────────────────────────
+   Prose measure and nothing else. The one page whose whole job is to say what the rest of the site
+   can and cannot prove is the wrong place to spend the accent on decoration, so it carries no
+   colour of its own — the only tinted thing here is a contact address nobody configured, which is
+   amber for the same reason every other unverified value on this site is.                       */
+
+section.about { max-width: 68ch; }
+section.about h2 { margin-top: calc(var(--cpad) * 2.5); }
+
+dl.identity { font-size: 13px; margin: var(--cpad) 0; }
+dl.identity dt { margin-top: 8px; }
+dl.identity dd { margin: 2px 0 0; color: var(--dim); }
+
+ul.sources { list-style: none; margin: var(--cpad) 0 0; padding: 0; }
+ul.sources li { border-top: 1px solid var(--line); padding: var(--cpad) 0; }
+ul.sources .name { font-size: 15px; font-weight: 600; text-decoration: none; }
+ul.sources .source-head { display: flex; gap: var(--cpad); align-items: baseline; flex-wrap: wrap; }