SellstampBlog
Notes on the decisions,
including the awkward ones.
Every non-obvious thing in this product was argued in writing before it was built, and the argument usually ends with what the choice costs. These are the ones worth reading outside the team.
- Engineering
A second currency is a price you typed, not a rate we applied
Multi-currency pricing was refused here twice for good reasons. It shipped when we noticed the reasons all rested on one unexamined assumption.
ADR-0170
- Engineering
A buy link is a URL, not a row
The obvious way to build shareable checkout links is a table. We built the opposite, and the reason is what happens to a link years after you send it.
ADR-0129
- Security
HTML and CSS yes, JavaScript no — and why that is not a compromise
Merchants asked twice for custom theme code, naming Shopify, Sellix and Tebex. That naming turned out to be the answer rather than an argument against us.
ADR-0141
- Security
Why exactly one fraud rule is allowed to block
Every fraud signal we collect is a sentence a human reads. One rule refuses a checkout on its own, and the reason is a clock rather than a severity.
ADR-0087
- Operations
The status page nobody automated
Ours has no probes and no health checks feeding it. A human types every word, and that is the feature rather than the shortcut.
ADR-0062
What this is, and is not
Each post is drawn from a decision record — the document written before the code, naming the problem, the options, the choice and its cost. The date on a post is that record's date, not the day somebody decided to publish it.
What this does not do
- No RSS or JSON feed. There is no endpoint behind one, and a link to a feed that 404s is worse than no link at all.
- No tag pages, archive or author pages. Five posts do not need navigation; they need to be five good posts.
- No guest posts, no SEO filler and no “ultimate guide to” anything. If we have not argued it in writing and shipped it, it is not here.
- No publishing schedule. A post appears when a decision was interesting enough to be worth reading outside the team.
Open a store and try it.
Everything on this page is in the product today. Connect a provider in test mode and put a real order through it before you decide.