New guy

[Herman]

Hi, I’m new to Solr but have been doing application development for a
very long time. I was hired to implement Solr, replacing many home
brewed SQL queries. It has been a lot of fun, as I’m sure all of you
are aware as it doesn’t take much to beat SQL queries.

I’m now two months in and have Solr up and running to everyone’s
pleasure, the trouble is, I happened upon SolrSharp first and these
are the underlying classes our stuff is built upon. It is my
impression that SolrNet has a lot more going for it; frequent updates,
this group, and so forth, while SolrSharp seems to have less attention
focused on it.

My question is this: am I misguided in my feeling that I should redo
my work on top of SolrNet? The primary driver in my decision is what
appears to be ongoing enhancement and communication; however, I would
like to know more certainly that the two libraries are more or less
doing the same things.

Can anyone shed some light on my question?

Thanks in advance, Herman

[Ken Foster]

As they are both wrappers around, and not a replacement for, Solr, I'm guessing they are doing more or less the same thing. However, it looks like SolrSharp is at Solr version 1.2 in release code and 1.3 in their trunk. SolrNet is at 1.4.1 and they are already talking about going to 4.0. SolrNet is definitely kept current with Solr. It also is built from the ground up with dependency injection, automated testing, etc. in mind. It's a nicely architected framework and worth studying simply for it's educational benefits. Obviously I'm a user of SolrNet and not SolrSharp so keep that in mind, however, the fact that SolrSharp doesn't support v1.4 of Solr is a deal breaker for me.

--
You received this message because you are subscribed to the Google Groups "SolrNet" group.
To post to this group, send email to sol...@googlegroups.com.
To unsubscribe from this group, send email to solrnet+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/solrnet?hl=en.

[[mRg]]

Likewise for me the support of Solr 1.4 features was crucial in our decision. Also we use Castle / dependency injection throughout our applications so it was a good fit.

[Mauricio Scheffer]

Hi Herman. I agree with Ken and Stephen. Both projects do pretty much the same thing, and it's quite simple really: generate and send xml to Solr, then parse back the results.

SolrSharp developers seem to have stopped working on it, but that doesn't mean that the project has to die. Since it was the first .net client it has quite a few users. In fact, there are several patches floating around the issue tracker, the mailing list and blog posts. IMHO someone should step up, fork the repository in github, or at least contact the original developers and convert the codeplex repository to mercurial, and then gather and apply these patches.

Since Solr is quite backwards-compatible, I don't think SolrSharp will *break* if it's used with Solr 1.4, or if it breaks it should be an easy fix.

I started SolrNet not because SolrSharp lacked any features but because I *really* didn't like its API, for example forcing you to use inheritance and it feels too Javaish for my taste. I also didn't like the internal structure of the code, it's too tightly coupled which makes it inflexible and unit-testing it very difficult.

Other people might prefer SolrSharp's API over SolrNet (for example if they come from Java).

Yet other people prefer to "manually" create the HttpWebRequests and deal with XML themselves.

Bottom line: pick whatever fits your development style better. This is not only about the visible API but about the implementation as well. Ask yourself: "if something goes wrong with Solr*, what would I do? Would I be comfortable messing around with this codebase?"

Have fun

Mauricio