The site-building jargon in tutorials, explained here

Page content is pre-written HTML files; the server just sends them to visitors as-is, without generating anything on the fly. This site is a pure static site: change content = change files.

Pages are assembled in real time on the server, usually with a database and backend code (PHP / Node / Java). More powerful but costlier to maintain and easier to attack.

Frontend is everything you see and run in the browser (HTML/CSS/JS); backend is the code and database on the server. A pure static site can have no backend at all.

Putting your site files on an always-online server so people can reach them via a domain is deploying; the place the files live is hosting.

A web address humans can remember, like xuliming.com. It’s just an alias; the real locator is the server IP that the domain resolves to.

The server’s street number on the internet, a string of digits (like 1.2.3.4). A domain’s job is to translate a memorable name into this number.

The process of translating a domain into an IP. Adding an “A record pointing at the server IP” in your domain console is exactly this configuration.

The DNS record that points a domain to an IPv4 address. Hostname @ means the apex domain, www means the www subdomain; the value is the server IP.

Points a domain at another domain (rather than an IP). Common when hooking up a CDN: let www point via CNAME at the domain the CDN assigns you.

How long a DNS record is cached. Changing DNS doesn’t take effect worldwide instantly — you wait for each ISP’s cache to expire per the TTL, usually minutes to hours.

To serve a public website from a server in mainland China you must register with the telecom authority and get a filing number. Unfiled domains resolving to a domestic IP get blocked.

A cloud vendor’s packaged product for individuals / small sites: fixed plan, bandwidth and traffic included, simple, cheap — better for beginners than standard ECS.

A web-based server manager: click to install Nginx, create sites, configure SSL and manage files — no need to memorise command lines. It does use a fair bit of RAM.

An encrypted channel for logging into a server remotely, port 22 by default. Connecting a terminal to a server to run commands goes over SSH.

The most common web server / reverse proxy software. It serves static files to visitors, configures domains, forwards requests and turns on caching. This site’s nginx-site.conf is its config.

The folder your site files actually live in (like /www/wwwroot/yourdomain/). When someone visits a URL, the server looks up the matching file in this directory.

Protocols for uploading local files to a server. SFTP runs over an encrypted channel and is safer; you can also sync a whole site at once with rsync / tar over ssh.

Linux’s timed-task runner. Daily backups, per-minute monitoring collection and periodic news pulls all rely on it to fire scripts on schedule.

The encrypted version of HTTP, with the padlock. A site without it gets an “Not secure” warning and a ranking penalty — now essentially a baseline.

What makes HTTPS work — it proves “this domain really belongs to me”. Let’s Encrypt offers free certs, valid 90 days, auto-renewable.

When a visitor types http:// or an old link, it 301-redirects to https:// so the whole site runs encrypted. Remember to enable it after installing the cert.

Caches your site across many nodes worldwide so users fetch from the nearest one — faster access, and it helps absorb traffic for the origin. Billed by traffic.

When a CDN node has no cache (or it’s expired), it goes back to your origin server once to pull the original file. A lower origin-fetch rate means better cache hits.

Lets the browser and CDN avoid re-downloading the same file for a period. A common setup: don’t cache HTML, cache versioned static assets for a long time.

How you get users fresh content after a redesign: append a version number to the file name (?v=2026.10.15) or change the file hash — the URL changes, so the cache is void.

All the practices that make a page easier to index and rank higher in search: titles and descriptions, content quality, internal links, structured data, speed.

When a search engine stores a page in its index so it can appear for a query. Not indexed = it doesn’t exist (as far as search is concerned).

The program a search engine uses to visit and fetch pages automatically (Googlebot, Baiduspider). robots.txt tells them what they may and may not fetch.

A plain-text file in the site root declaring which paths crawlers may or may not fetch. A single wrong “Disallow: /” can shut the whole site out of indexing.

A file telling search engines “here are my pages and when they were updated”. Submitting it to Baidu / Google speeds up indexing.

Declares “the official address of this content is this one”, consolidating weight from duplicate pages (with parameters, multiple paths) onto a single URL so you don’t compete with yourself.

Tells search engines the relationship between language / region versions of the same content. A bilingual site uses zh-CN / en / x-default, all pointing at each other.

404 is the correct “page not found” status. A soft-404 is when the page is actually gone but you return 200 (e.g. redirecting home) — it misleads search engines, so return a real 404.

301 is permanent and passes weight to the new address; 302 is temporary. Forcing HTTPS and moving old links to new pages generally use 301.

Marking up content in a fixed format (schema.org) so machines can read it as data (tutorial steps, Q&A, breadcrumbs) — it can show rich-result styling in search results.

Low-volume but very specific queries, like “aaPanel port 8888 won’t open”. Little traffic each, but a lot in aggregate and highly targeted — the words a new site should chase first.

Links between pages on your own site; anchor text is the clickable words. Good internal links let crawlers traverse the whole site, push weight to deep pages and tell engines what a page is about.

A lightweight data format (key–value pairs). This site’s scenarios, steps and FAQs all live in .json files. Strict format: double quotes, no trailing commas.

Browser local storage that survives closing the page. The wizard uses it to remember your checkbox progress and theme — no account, no backend.

How the frontend requests data from a URL. This site’s daily and papers pages fetch external APIs live and render them. Cross-origin calls require the other side to allow CORS.

For security, browsers by default forbid page A from reading site B’s data unless B explicitly allows it (returns access-control-allow-origin). The “cross-origin error” from hitting an API directly is this.

A lightweight syntax for writing formatting in plain text (# heading, - list, `code`). Blogs / docs often write in it, then a generator converts it to HTML.

A backend-free comments solution that stores comments in GitHub Discussions. Good for static blogs — no comment server or database to maintain yourself.

Free, static-site-friendly hosts: push code, auto-deploy, HTTPS included. Personal blogs / docs can start at zero cost (mainland access speed varies by network).

Earning a commission when you use your own promo link to guide users to place an order (Cost Per Sale). Cloud vendors and e-commerce networks offer it.

Two ad-billing models: CPC is cost per click, CPM is cost per thousand impressions. Used to estimate revenue when monetising traffic.

A rel attribute on promo / affiliate / paid external links, clearly telling search engines “this is an ad link, don’t pass weight”. Both a compliance requirement and an anti-spam signal.

PV is how many times pages are viewed; UV is the count of unique visitors. One user browsing 5 pages = 5 PV, 1 UV — the basic metrics of site traffic.

The share of visits that complete the action you want, e.g. read a step → clicked a promo link. This site’s conversion tracking quantifies that ratio, not just PV.