The usual way is to create an index on generation time and serve it statically. JS just uses the index to do the search. It's a big file, so I'm not saying it's a great solution for everyone, but it works reasonably well.

Of course, for my site I just redirect the user to a search engine plus `site:stavros.io`.

See https://gohugo.io/tools/search/. Not sure how well they scale to thousands of posts, but they work by statically generating multiple static search index files at build time that are queried via client JavaScript when hosted. The search UX is actually really good because they tend to respond instantly as you type your query and allow complex queries.

Do you happen to know if any of those support faceted search (ie searching and filtering by date, category, etc)?

I've used https://lunrjs.com/guides/getting_started.html briefly and it has lots of options for things like different fields, complex queries, fuzzy searching and wildcards. Didn't notice anything specific about dates but you could always add date as a field then filter out a date range manually at the end. I'm sure there's better libraries now as well.

We've gone from SSGs for ease, speed and reduced resources, to talking about implementing search with multiple megabyte client side indexes and hundereds of thousands of prerendered search result pages.

When does this become 1 step forward with the SSG and 2 steps back with search solutions like this?

You don't pre-render the search pages. You generate some search index files on the build step (something like a map of keywords to matching post URLs), and then client side JavaScript requests the search index files it needs on demand and generates the search results on the page. For a modest blog, I think the compressed index can be a few 100K. A single large image can be bigger than that.

Nothing is perfect, but the above is really simple to host, is low maintenance, and easy to secure.

This is just server side search with more steps, since the index needs to be selectively split and returned, and then search results page generated from the index.

Sorry, I don't mean to come across as disagreeable. You're right, nothing is perfect, and this is obviously a workable and usable solution. My issue is if we analyse it beyond "it looks like it works", it starts to look like a slightly worse solution than what we already had.

Nothing wrong with moving backwards in some direction, as long as we can clearly point to some benefit we're getting elsewhere. My issues with SSGs is most of the benefits they offer end up undermined if you want any interactivity. This is a good example of that, as you end up compromising on build time and page load time compared to an SSR search solution.

> it starts to look like a slightly worse solution than what we already had

You don't see how the server based solution is an order of magnitude more effort to maintain, monitor, optimize, and secure compared to hosting static files? Especially if search is a minor feature of your site, keeping the hosting simple can be a very reasonable tradeoff.

Lots of blogs/sites also do fine with only category and tag filtering (most SSGs have this built in) without text search, and users can use a search engine if they need more.

My point the entire time has been that an SSG is not good for sites that want interactivity.

I agree with you, lots of blogs/sites do fine with just tag/category. Those sites don't have interactivity and so you could SSG them and use free static hosting. We're in agreement on this. But I have explicitly been talking about sites that want interaction. Still - how many personal sites that "do fine without interactivity" are actually "interactivity was too much of a technical/maintenance challenge"?

I completely see how a server based solution is an order of magnitude more effort. We agree on that too. Running any servers efficiently typically has fixed overhead for server infra and processes. Running 3 pieces of server software or even 3 servers is harder than 1, but significantly less than 3x harder.

If you want interactivity, a server is needed. This means you are already maintaining, monitoring, optimising and securing a server solution. Adding SSG processes alongside your existing server solution does not remove any problems relating to servers, because you still have servers. It does add complexity though because you're running servers, and some SSG pipelines. These architectures clash and create more work. Running 2 conflicting architectures can be a lot more more than 2x harder than running 1 architecture.

If you want interactivity and don't want to run a server, then you must use a 3rd party service. This brings its own issues, especially if you have a personal website because you want to own your own stuff.

I feel like you're repeating to me "servers are hard, SSGs are good" - you can't have an SSG without a server, and you definitely can't have interactivity. If you have a server anyways for comments or search, then by using SSGs, you're not removing your need to deal with servers, you're adding an extra picky new thing to your servers. And what complex feature of a personal website does the SSG solve specifically? Generate the text/HTML. Compared to comments and search, generating html from markdown is so painfully easy. SSGs tie our hands behind our backs for the hard interactivity problems to solve the easy text rendering problem.

This seems like a really, really crystal clear point to me. But you, Jeff and Susam have all tried to say that SSGs are generally better, easier, safer or faster. I felt like that from 2016 - 2024. But when you've been pressed, all of you have ended up saying variations on:

- SSGs are great for sites without interactivity (agree, you know and I know though, most personal websites do want some form of interactivity, even just search or comments, and the only reason not to is technical hurdles.)

- Comments and Search are doable with SSGs, but not trivially, or without compromising on things which were traditionally seen as crucial for an open and accessible web (agree, see jeffs comments and search tickets)

- You use SSGs not to solve any particular problem, but just for fun. (agree, that's fine, but if that's your only reason, that's probably a sign there's virtually no technical advantages.)

- The problems you used the SSG for originally have been solved with compute power, storage, connectivity and bandwidth becoming more available over the past decade

I really appreciate and respect your's, Jeff's and Susam's time to respond, I want to be in agreement with you, as you all have the experience and reputation in personal websites. I just can't see it though. I mean this whole HN topic is about Jeff's post "Jeffgeerling.com has been migrated to Hugo" which on the surface looks like a success for the ease of SSGs, but when you probe deeper, it's the text that's been migrated, and comments and search have been disabled.

Why are we here if SSGs are so obviously the easiest, most reliable way to run a personal website? Jeff's site is currently working less than it did before and the fixes are future TODOs are essentially "pick a compromise" (sorry, no offence Jeff - I still think you're great!)

> But you, Jeff and Susam have all tried to say that SSGs are generally better, easier, safer or faster.

I'd just like to clarify that this is not what I was trying to say. In fact, once a website needs server-side programming to support interactivity (comments, subscription, etc.), I don't think SSGs are necessarily better. I think we both agree there to an extent. I think there are tradeoffs.

There is a certain inherent complexity in addressing concerns like performance, portability, etc. that, as waterbed theory suggests, just gets pushed to different places depending on whether we choose an SSG-based solution or non-SSG-based solution. What I have been trying in my other comments is to explain my reasons for why I made the arguably questionable choice to use an SSG for content pages while relying on server side programming for comments and subscriber forms.

I certainly see the appeal of serving the entire site, including both content pages and comment forms, using server side programming. It does simplify the 'architecture' in some ways. However, since I maintain my personal website as a hobby, I need to make the tradeoffs that feel right to me. I realise this is not a very satisfying answer to the question of whether a hybrid solution like mine (SSG plus server side programming) has merits. But I was never trying to argue its merits in the first place.

This is all entirely fair, thank you, I understand, I didn't mean to misrepresent.

> However, since I maintain my personal website as a hobby, I need to make the tradeoffs that feel right to me. I realise this is not a very satisfying answer to the question of whether a hybrid solution like mine (SSG plus server side programming) has merits. But I was never trying to argue its merits in the first place.

Hey, I'm a firm believer in the idea that the best solution for a personal website is the one that's the most fun for the owner to maintain! The reminder that it's not all about technical merit makes yours one of the more satisfying answers - thanks again for the considered responses:)

Sounds like we agree! There's no perfect solution and I'm not saying SSGs are always the best fit. It really depends on the project and what the people involved value.

For my own solo projects, the ease of hosting for static sites is often such a big win for me that I'm willing to forego some interactivity even though more interactive features would be nice. Knowing that I can upload several static sites and they'll run themselves without any maintenance for years and without potential security problems keeps my life nice and calm. It depends what you want to prioritize.

I agree, nothing is perfect, there's a time and place for everything - I just think the coverage and advocacy of SSGs is disproportionate to the number of places where they improve things. I'm going to summarise the discussions I've had in this topic in a post over the next few weeks and I'll post it here. Thanks for humouring and challenging me :)

I still do feel there's something not quite right. Hopefully I'll be able to get it out in the summary as I think we're all done with walls of text here. But I think you actually captured it perfectly in your last remark:

> For my own solo projects, the ease of hosting for static sites is often such a big win for me that I'm willing to forego some interactivity even though more interactive features would be nice.

They make the web worse for the world by tempting the developer to take the easier, less interactive route than what would have been taken in the pre-SSG world.

I think I'm troubled by this because the idea "it makes it easy to do the lazy, worse thing" seems completely at odds with the fact that it seems to be what the people I would consider leaders in the field are doing. That's why I still wonder if this view might be a me problem, or maybe the field is just moving backwards? I don't know.

Anyway, thanks again!

> They make the web worse for the world by tempting the developer to take the easier, less interactive route than what would have been taken in the pre-SSG world.

If I didn't have the static site option though, I might not host a project at all because I don't want to have to deal with the maintenance burden. I haven't kept up but it would be nice if there were more options between a no-maintenance static site hosting, and a high-maintenance dynamic site.

Would be amazing if the web/browsers had standardized ways by now of easily adding common website features like authentication, search, comments, votes, and saving user data (via bring your own storage) that didn't require everyone to host and maintain a custom stack. Email is a solved problem that you don't have to host yourself for example, why is it still such an effort for the other common website features?

> I haven't kept up but it would be nice if there were more options between a no-maintenance static site hosting, and a high-maintenance dynamic site.

This is essentially my position in fewer and less argumentative words, including the "I haven't kept up bit", really:)

Your last paragraph was also really thought provoking. I'm torn on that though. I agree it's an almost off-putting amount of time for anyone who wants a personal website to reinvent the wheel when the client browser could help.

The main issue with that for me is that browsers require organisations to build and maintain, so I think we should always be careful about removing responsibilities from website creators and giving it over to browsers.

I guess the SSG and the browser points are both freedom vs convenience type tradeoffs where everyone has their own personal preference. It's all Web stuff though, so we should probably try keep it all open, flexible and resilient IMO.

Interesting attempt at bad faith discourse.

Assuming 500 bytes of metadata + URL per blog post, a one megabyte index is enough for 2000 blog posts.

As already mentioned, you don't generate search result pages, because client side Javascript has been a thing for several decades already.

Your suggestion of converting markdown on every request also provides near zero value.

Writing a minimal server backend is also way easier if you separate it from the presentation part of the stack.

Based on https://news.ycombinator.com/item?id=46489563, it also seems like you fundamentally misunderstand the point. Interactivity is not the point. SSGs are used for publishing writing the same way PDF is used. Nobody sane thinks that they need a comment section in their PDFs.

I'm sorry, I find your post insulting. I wasn't engaging in bad faith discourse intentionally, but your assumption that I am - ironically - doesn't feel like appropriate etiquette here either. Despite that, I'll try answer sincerely.

Your 1 megabyte index file has just added over 2 seconds to your page load time in 30 different countries based on average internet speeds in 2024. Chuck in some pictures, an external comment library and your other SSG hacks, and you've just made your website practically unresponsive to a quarter of the planet and a bunch of other low powered devices.

Value is relative. The benefit of rendering markdown on every request is it makes it easier to make it dynamic, so you don't need to do SSG compromises like rebuild and reupload multiple pages when a single link changes.

You're replying in my thread here, to my original points. My original points were that SSGs don't make sense for sites with interaction, which is why were were discussing the limitations of SSG search approaches.

> SSGs are used for publishing writing the same way PDF is used. Nobody sane thinks that they need a comment section in their PDFs.

Thank you! We're in agreement, it doesn't make sense to use SSGs for sites that require interaction. When you do, it forces the rest of your site to do the compromising search stuff like we're discussing here.

> Your 1 megabyte index file has just added over 2 seconds to your page load time

It might not be intentional (I doubt), but your replies really read like bad faith discourse.

> there is a Drupal 7 site that I need to migrate to something

You may be interested in Backdrop, which is a maintained fork of Drupal 7.

https://backdropcms.org/

(No experience with it personally. Only know about it from a friend who uses it.)