[tw5] Re: The Performance At Scale Bummer ... Initial Notes

Josiah, As you say "Whilst 8,000 Tiddlers of Tweets with tags that rely on filtering gets very slow" when looking at the specific case I realised the main thing searched for was the tweet, so in a tiddlywiki for this purpose why is it a secondary fields when it could be the title. Then in None

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

Could subvert the minimum length field. Put this below first div in $:/core/ui/SearchResults: <$reveal tag="div" state="$:/config/Search/MinLength" type="nomatch" text="50" default=""> <$button>Hide<$action-setfield $tiddler="$:/config/Search/MinLength" text="50"/> and replace contents of

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

> > Yes -- that's what I said. But is there a setting to turn off key-by-key > searching and replace it with traditional type-and-enter ? > > Thanks! > Mark > Hi Mark, couldn't that be done with a <$keyboard key=yourkey actions="set statetiddler on"> before the search input field ... and

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

Yes -- that's what I said. But is there a setting to turn off key-by-key searching and replace it with traditional type-and-enter ? Thanks! Mark On Tuesday, April 10, 2018 at 8:41:42 AM UTC-7, PMario wrote: > > On Tuesday, April 10, 2018 at 5:35:07 PM UTC+2, Mark S. wrote: >> >> Since there's a

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

On Tuesday, April 10, 2018 at 5:35:07 PM UTC+2, Mark S. wrote: > > Since there's a hidden setting to control minimum key length, it must be > possible internally to change the mechanism so that it only searches upon > enter. Maybe there could be a hidden setting to turn off key-by-key >

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

In the past when I've tried the type-ahead setting, I couldn't find that it made any difference one way or the other. I'll try again. What I mean about the database on the backend, is that big companies like Google and Amazon use some streaming/buffering Ajax technology where search key-by-key

Re: [tw5] Re: The Performance At Scale Bummer ... Initial Notes

> On 10 Apr 2018, at 14:09, PMario wrote: > > VERY useful info! Who knew but you? :-) the problem with Hidden Settings is > they are ... Hidden :-). https://tiddlywiki.com/#Hidden%20Settings Best wishes Jeremy -- You received this message because you are subscribed to

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

On Tuesday, April 10, 2018 at 3:08:04 PM UTC+2, PMario wrote: > > On Tuesday, April 10, 2018 at 2:42:40 PM UTC+2, @TiddlyTweeter wrote: >> >> VERY useful info! Who knew but you? :-) the problem with Hidden Settings >> is they are ... Hidden :-). >> > > For a reason. ... They have the ability to

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

On Tuesday, April 10, 2018 at 2:42:40 PM UTC+2, @TiddlyTweeter wrote: > > VERY useful info! Who knew but you? :-) the problem with Hidden Settings > is they are ... Hidden :-). > For a reason. ... They have the ability to cause "strange behaviour" if used wrong. -m -- You received this

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

PMario VERY useful info! Who knew but you? :-) the problem with Hidden Settings is they are ... Hidden :-). PMario wrote: > > There is a HiddenSetting >

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

On Monday, April 9, 2018 at 5:21:21 PM UTC+2, Mark S. wrote: > > I perceive the results-as-you-type thing as more of a toy -- not a > usability development. > You can change it. There is a HiddenSetting

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

How about a "don't do it till you click" button? On Tuesday, 10 April 2018 07:12:22 UTC+2, Mat wrote: > > Apropos results-as-you-type, I just (re-)stumbled over this, for anyone is > interested. Who knows, maybe it has some superior algorithm for big data > sets or some other useful aspect that

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

Apropos results-as-you-type, I just (re-)stumbled over this, for anyone is interested. Who knows, maybe it has some superior algorithm for big data sets or some other useful aspect that we can use: Awesomplete <:-) -- You received this message

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

I perceive the results-as-you-type thing as more of a toy -- not a usability development. Not just TW, but anywhere it's used. It may make sense when it's attached to a powerful back-end server database, but not so much when you have a local, non-indexed data source. In the case of TW, it

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

Ciao Mark S. I guess part of the issue is that "usability" developments can have consequences unforeseen at the time??? J. Mark S. wrote: > > > Such as a search that responds keystroke by keystroke. > > I would love to turn that feature off. You don't need 85,000 tiddlers to > have it slow

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

Ciao TonyM & all interested I do think that a "Rules Of Thumb" note on SCALABILITY could be very useful. The issue is expressing it correctly. So much of performance is an EMPIRICAL issue, that without testing variant use cases, you can make TW look worse than it is. That is not a good idea.

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

> Such as a search that responds keystroke by keystroke. I would love to turn that feature off. You don't need 85,000 tiddlers to have it slow things down -- even a few hundred is enough on a tablet. I'd much rather type and enter without the performance and interference hit. -- Mark --

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

Josiah, I have given this some thought, and have a lot to add. Perhaps we should start a discussion specifically on settings and tools for large wikis? I just loaded the thousands of standard tiddlers (mostly tweets) from the large TiddlyWiki exhibiting performance problems. When it come to

[tw5] Re: The Performance At Scale Bummer ... Initial Notes

Josiah/Jed, Has anyone tried to see if noteself with its in browser counchDB can perform faster with large wikis? Tony On Saturday, April 7, 2018 at 7:48:55 PM UTC+10, Jed Carty wrote: > > Looking through my things, despite using the Object.keys().forEach > structure many times there is only