Searching the Internet or specified URLs
SV offers you the option of searching the whole Internet, or searching a URL of your choice.
It also offers you the option of selecting one of the collected texts on our site.
The Internet search option is just like a standard Internet search on one of the usual search engines; the only difference is that you see each record displayed as a record visualisation, as if someone had printed out the record and highlighted the places where your keywords occur.
The "Search URL" option is useful if you want to search a particular set of records.
We don't encourage people to do repeated SV searches on online archives, for reasons of netiquette
What a lot of users will find useful is that if they have their own set of documents on their own website, they can use the SV to search those documents.
For instance, if you have a particular interest in Shakespeare, then you can download his works from a site such as Project Gutenberg, put them on your website, and search them at your leisure.
The collected texts on our site are mainly there as examples of what SV can do – we've included some essays about ways of using SV on large texts.
We've included some Shakespeare texts, which demonstrate some interesting patterns in his use of language.
We've also included some historical records which are a collection of documents written by different people, as an example of how you can use SV on very large documents.
Number of keywords
You can choose up to five keywords for an SV search.
These can be text, numbers or a combination of both
The current version treats numbers as if they are a word.
We've used five as an upper limit mainly because most people find it hard to interpret a record visualization with more than five colours in it.
Another factor is that increasing the number of keywords slows down the search.
We've found five keywords to be the best tradeoff for most people.
You can search for a phrase by putting those words in inverted commas.
For instance, if you want to search for the words empire builders in that precise phrasing, then you can type in the keyword phrase "empire builders" in inverted commas.
That will find any records which contain those two words next to each other in that order; it won't show you results for "builders of empire" or other variants on the original phrasing.
If you do this, then the entire phrase counts as a single keyword.
You can search for synonyms on the SV by separating the synonymous words with a comma (without spaces).
You can treat any words you want as synonyms of each other.
For instance, you can treat dog and cat as synonyms of each other by searching for dog,cat.
You can combine this with other search words.
For example, you could search for dog,cat photo and this search would produce results for photos of dogs and/or cats.
This can be useful when searching for documents about people who are known by several names, or for comparing documents that use different names for the same thing.
Each group of synonyms will count as one keyword.
Searching for hits within a word or whole-word hits
Sometimes it's useful to be able to search for a hit within a word: for instance, if you search for the keyword sign as a within-word hit, that will treat signage and signing and signs as hits.
For some searches, though, it's necessary to go for whole-word hits.
For instance, if you're looking through Shakespeare's texts for gendered language, then you might want to search for he and she to see how often each word occurs.
You'd need to search for them as whole-word hits, since if you did the search as within-word, the word he would also get hits for her and the and for she, producing misleading results.
Number of records per screen
You can choose to display different numbers of records on the screen at a time.
If you select a higher number, then you can see more records at one time.
This is useful if you're dealing with short records.
If you're dealing with longer records, then you'll probably find it easier to display fewer records on screen at a time.
The software will automatically adjust the display to make each record visualisation as broad as possible, which means that you'll be able to see more of the record without scrolling down.
It's sometimes useful to select a higher number of records per screen even if that means scrolling.
The reason is that if you're looking for structures within a text (e.g. whether it's broken down into sections with different themes) then you'll often find it easier to do this with comparatively narrow record visualisations.
You can select different square sizes for the squares within the record images.
This is useful when you're dealing with large documents.
If you want to see the whole of a large document on a single screen, then selecting a smaller square size will help.
As a rough rule of thumb, on the smallest square size, you can see a document two or three hundred pages long on a single screen.
At present, SV comes with pre-set colours for the keywords: the first keyword you select is represented by red squares, and the second keyword is represented by green squares, for instance.
We've pre-selected these colours and their shades to be distinguishable by people who are colour blind.
If you're searching for keywords which are names of colour (e.g. "yellow hair") then it's confusing to see the word "yellow" represented by a red square.
(The technical reason for this is the Stroop effect.)
One way of handling this is to choose the order of your keywords, and/or to add some gibberish words which won't get any hits, so that the keyword colour matches the meaning of the word (e.g. making sure that "yellow" is your fourth keyword, so it is represented by yellow squares).
You can save screenshots of searches using the relevant option or utility on your computer (most machines have a key combination which does screenshots).
If you want an image of a particular search visualisation, then you should click on that visualisation; this will bring up a big image.
You can then use the "save image" option on your computer (a right-click on the mouse for most users) to save that image.
The image you get this way is high resolution and easily editable.