Back to field notes

Why a market desk should start read-only

Separating observation from execution makes the first version of a market workstation clearer, safer and easier to validate.

FR

Why a market desk should start read-only

A desktop market tool becomes consequential the moment it can place an order. That is not only a new button: it changes the security model, test surface and way every state must be communicated. IBDesk is deliberately beginning on the other side of that boundary.

Observation is already a complete problem

Quotes do not arrive as one perfect stream. Top-of-book data, five-second bars, tick-by-tick trades and market depth can have different permissions and different failure modes. A useful desk needs to show which feed is live, frozen, delayed or unavailable without turning every limitation into a generic error.

That is why IBDesk keeps subscriptions scoped to an instrument and a feed. If depth is unavailable, the quote can remain useful. If the broker reconnects, the desk can restore the feeds the workspace actually requested.

Local context matters

The current foundation archives typed market data into daily local SQLite files. This supports historical repairs today and lays groundwork for operator-visible replay later. The archive is not presented as a perfect exchange record: completeness and provenance still matter.

A clear boundary earns trust

The first public milestone will not imply that account management, order entry or complete after-hours coverage already exists. A smaller honest surface is more useful than a broad list of unfinished claims. Execution can be designed later as its own reviewed product boundary.

IBDesk requires a separately installed TWS or IB Gateway session and the appropriate market-data permissions. It is independent software and is not affiliated with Interactive Brokers.