"parallel" database discussion

[Robert Lauriston]

Jim Leff and Melianie Wong were kicking around the idea of building some sort of off-site "parallel" database as a "creative workaround: for the loss of the old restaurants database:

http://chowhound.chow.com/topics/8274...

"Can we PLEASE not discuss this in this over-stuffed, rambling thread? PLEASE someone start a new one."

Here you go.

[Melanie Wong]

Thank you, Robert. What I see is a salvage operation to save off a snapshot of just a small portion of the existing database before the pages are retired. Once all the data from the vendor have been removed from Chow to comply with its contract, the data we’ve saved and cleaned up could be imported back into the system. I agree that any database needs to be linked to discussions on the board to be valuable.

Critical to this is to leave the current programming and software in place that supports the restaurant database until such time as Chow can develop a replacement. Meredith has said that the database is diseased and needs to be put down. She also said that it was slowing the performance of Chowhound. But I’m wondering if the amount of data associated with the program slimmed down from 900K restaurant records (aka IDs) to somewhere between 0 and 100K, would it work well enough. We’d be starting from zero. Users of the site would be able to manually enter restaurant data, just as they can now.

Meanwhile, depending on what can be salvaged from the existing dataset, we’d be working on filling in the holes in each record. Manually if we have to. In another thread I’d asked for the names of the restaurants that I’d linked to and was told it was not feasible. Instead of that data crawl, I’d like to see an export of the restaurant IDs that have been linked to a discussion. That would be the records that pop up under “most discussed” in the restaurant database search function. These are the places that chowhounds most care about and should be the core of our new database.

A couple days ago I took a look at how many that might be for selected regions. This is the number of restaurants linked/number of restaurant records in the region.

4739/11515 – Manhattan (41%)
8469/21245 - SF Bay Area (39%)
7792/39415 - Los Angeles (20%)
1575/7713 – Phoenix (20%)
3715/16836 – Boston (22%)
3716/25131 – Toronto (15%)
1610/19826 – Chicago (8%)
921/8734 – Seattle (11%)

It was interesting for me to see the difference in behavior across regions. Anyway, of the overall database of 900K, maybe 10 to 25% would be saved.

We’ve been told that address information and websites input to the database cannot be exported due to legal constraints. The export files we received had the restaurant names, so presumably that is fine. I’d like to know what other data in each restaurant ID can be exported and if board region can tagged to them.

[rworange]

Melanie ... give it up. it is dead. i know as of this post I have.

If they are removing the database in March, all the coding has already been removed and they are in the testing phase.

Chow is all ok directing traffic to other sites should anyone want to put that effort into it.

A little crystal ball here ... in the unliely event it gets rebuilt no on will look at what was good about the current database or even ask any of the people who put a lot of work into this one. It will be business as usual at Chow. Make the decision and let posters eat it.

[mudaba]

I want to be really clear on this: CHOW supports people linking to Yelp, Google +, and especially to the restaurants themselves, etc, for address information. It is what we are going to be doing in our stories and other places on the site as well. We cannot support anyone taking information from this site (the exact info we are taking down because we do not legally own it all), and trying to create a new site out of it. We're going to lock this thread, and would ask that it not continue in the previous thread.

Thank you!

[mudaba]

Hey everyone,

I wanted to just weigh in on this thread and to ask that everyone have a little patience. We need time to plan any new restaurant database solution. Rushing it right now, in the middle of trying to clean up the last, poor build is not the way to get a good solution. So for now, if you want to link to Yelp, it's fine. Perhaps once we have a chance to reevaluate and build something truly worthwhile and undiseased, you will find that your needs are met. But it takes time.

An announcement that I just made on another thread is that a longtime Chowhound is coming to work at CHOW in March. His user name is Dave MP. I am very excited to have him join us here, and to digging into questions of how to build a strong mobile site for Chowhound, how to get address information onto the site, and how to build features that really work. Dave MP has been active on the site for over 11 years, so you can trust that he will have your best interests at heart. We will also be user testing larger new features like restaurant address linking, but again, it takes time. Please look at the positive changes we are making and understand that it is a commitment to good things in the future for Chowhound.

Thanks, Meredith of CHOW

[rworange]

Meredith,

I know you are one of the good guys. However, a workable interface to the existing database meant five years of waiting with zero action. I won't hold my breath waiting. However, if it happens it will be a nice surprise. In the mean time I would like the export file of my Guatemala information. If that can't happen then I'll just cut and paste all that info so I can put it elsewhere.

[Melanie Wong]

Meredith, thanks for the public announcement about Dave MP and his new role. I'm happy he'll be back with us in the Bay Area, and even happier that he'll be part of the Chow team. Great hire.

We know it was a poor build. But please leave the existing programming up that supports the restaurant database. Wipe out all the data and dump the existing pages, but leave the functionality for people to start again until there's a replacement. That gives you an incentive to build a replacement.

[Robert Lauriston]

The main potential obstacle I see would be CBS not allowing links to the other site.

[rworange]

Well, I'm currently linking to yelp now for that info. It is not in the rules that we can't. I enter a review or a response with information and then link to yelp for the address information.

I would hope that CBS would not object since it doesn't want to provide that data on this site. is cutting and pasting the info from yelp or a parallel database any different than just providing an easy link to the actual info.

Is it the intent of Chow or CBS to make this site as difficult to use for posters as possible.

There is not a good restaurant database on the web for a lot of reasons and a paralllel database is a market waiting to happen. However, had Chow/CBS saw the potentional of the current database that would have been a lot better. No one, including me, wants an outside source.

[hyperbowler]

Yes, I might have missed this in the parent thread, but I'm surprised that CBS hasn't officially suggested what should replace the restaurant database in terms of providing addresses in threads, hours, etc. I think this decision should ultimately be up to the users, but they have to have a preferred method, and I would assume one that doesn't involve traffic to their competitors.