Google Custom Search not as cool as Rollyo.com

happygeek 0 Tallied Votes 277 Views Share

Google has launched a new search engine, or rather has made the old one available with a few user configurable tweaks. The (CSE) is now open for business and provides a brand new and tailored search experience, or at least that is what Google is saying.

It still uses the Google core search technology so there are no nasty surprises when it comes to results, only good ones. Because by using Google CSE, the user chooses to prioritize or restrict searches based upon the sources, websites and pages, you specify. In other words you get to build your own Google index, and search just that instead of the whole shebang.

You can even share your index and let others contribute to it if you like.

I like the ability to be able to give selected pages and sites a higher priority and ranking within the larger Google index, that’s both neat and useful. I like the bit of code you get given so you can have your own Google CSE search box on your site or blog. I even like the concept of creating a custom search tool that can then earn me money by using the AdSense program alongside it.

But you know what, it isn’t exactly a brand new idea and I think I prefer the approach taken by with their Roll Your Own search service based on the Yahoo engine. This brings search customization down to a per domain level, narrowing your search sources to a list of specific sites that you know and trust. These can then be published as ‘Searchrolls’ to allow others to share the same resource.

What’s more if you like a Searchroll not only can you use it, but you can edit it to make it yours without changing the original. And, you can have as many Searchrolls as you like, one for each area of research for example. You don’t even have to roll your own from scratch, search for an existing one and customize that. The only downsides are restrictions to only 25 sites per Searchroll, and the fact that if you don’t like Yahoo searches (and Feedster which drives the blog search element) you are pretty much stuffed of course.

Dani AI

Generated

Good back-and-forth so far. The thread effectively illustrates two different design choices for custom search: a tightly curated list of trusted domains versus a broader, tunable engine that searches a larger index. favors the curated approach for control and predictability; plans to experiment with the other model. Which is “better” depends on what you need: strict precision and known sources, or broader coverage with less manual curation.

Practical checklist for deciding and deploying either model:

  • Define the goal and success metrics first (what answers must appear, what can be missed).
  • Create a representative test suite of 20–50 real queries and use it to compare result precision and recall.
  • Build small pilots: one focused curated list and one broader engine with tuned weights; embed both behind a simple UI so you can switch and compare.
  • Design the UI to show provenance (where a result came from) and offer an explicit “search the wider web” fallback when the curated set returns poor coverage.
  • Put a maintenance plan in place: review query logs, add/remove sources, and re-run the test suite on a monthly cadence.

Quick tuning and troubleshooting tips: monitor zero-result and no-click queries to spot gaps; avoid single-site dominance by promoting representative pages rather than entire domains; when community contributions are allowed, require a lightweight moderation step and keep a changelog of edits so you can revert harmful inclusions.

A hybrid approach often wins: use curated lists for high-value verticals and provide an opt-in web-wide fallback for discovery. Start small, measure with your test queries, and iterate based on real usage data.

Dani 5,664 The Queen of DaniWeb Administrator Featured Poster

I'm really liking the Google custom search engine, and will be tinkering with it today.

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.