Home | | Privacy Statement | Advertisers | Contact Us

The Letter Repository is an experimental way to fund the digitalisation of history through the use of funds generated by advertising. Therefore there is a direct relation between the amount of traffic this site attracts and the ability to add to our database. Here we explain our service, how our site is structured and information on how to best categorize our material, so that fair and interested netizens can be directed to content they want.

domains & hosting

The Letter Repository is a stand alone service, has all 100% original content and is independent of any other website or service. However due to the way it grew and our desire to have memorable domain names, our service is somewhat fragmented across several domains. Here we list all the domain names we own.

list of domains
letter.ie - This site and the main database.
war-letters.com www, civil Access to former content and military conflicts and some subdomains
letters.ie love, ww1, ww2 Subdomains containing category pages.
letterrepository.com - Parked domain pointing to our homepage.
letterrepository.org - Parked domain pointing to our homepage.

Search engines may consider these sites to be of the same service and links between them to be internal links. All domains presently hosted by Digiweb Ireland at the IP

TLD ".ie"
Our main service is accessed through the url "letter.ie" with the tld of "ie" which is associated with Ireland. This along with our hosting provider also being located in Ireland may falsely indicate that we are providing a service exclusive or at least predominately for an Irish audience. This is not so.

As of the November 2009 only one of the thirty-seven lots (3%) and with less than 5% of the total letters are addressed to addresses in Ireland. We are not ignoring Ireland but have decided to focus on the English speaking world of which approxiately 5% is Irish. Approx. 68% of our letters are addressed to the U.S. and 50% of our current traffic. Other well serviced countries are the UK with approx. 25% of letters.

In the future we hope to add a lot more languages to our service. We have considerable material in our procession written in Latin, Spanish and French which will be added when funds allow.

structure & layout

Note that each letter belongs to a lot which is a group of letters just the same as were obtained by us at Letter Repository. We did not attempt to sort these letters as to do so may destroy some of their history. Lots are however normally to the same recipient(s), with multiple letters in a lot being from the same sender/author. Contained in the url of each letter in this database is two four digit numbers which uniquely reference it. The first four digits identify the lot and the last four the specific letter in that lot. Letters in lots are loosely arranged in cronologically order.

A short description of each lot, plus a link to all the letters in a lot are accessible through the 4 digit lot number. You should not rely on this summary as it contains a very short snippet of the content of the letters and may indeed contain factual errors. The front of each envelope is accessible by the two 4 digit index numbers seperated by a backslash with ".jpg" appended if it exists. Approx 20% of letters had no envelopes.

Each page which makes up a letter is accessible through the 8 digit index number plus "p1.jpg" for the first page, "p2.jpg" for the second page and so on, appended. Likewise photos or other attached documents enclosed in an evelope is accessible by index number plus "a1.jpg" and so on appened. Occassionally images and other documents (not letters) were included with the group at the point of purchase or obtainment. In these cases the images of the documents are available from the root of the lot directory. These are normally given the filenames "d1.jpg", "d2.jpg" etc.. for documents and "i1.jpg", "i2.jpg" for photos and images. A link to all these is available from the lot page.

As an essential part of the digitalisation process, the content of each letter is transcribed into ascii text form. This allows computers to read the content of the letters and through the use of third party translations services (currently google translate) allows the words to be translated into serveral world languages. This can give a new perspectice to a monolingual student or researcher. As part of this transcribing process many mistakes are made in the first transcription of a letter due to hard to read handwritting and/or lack of time by the transcriber. Therefore we allow our visitors to edit our letters. A list of pages flagged as under-transcribed or un-transcribed is available here.

All this makes our content as accessible to as big an audience as possible. An edit button in the top left corner of the text area for each page of each letter is provided. By clicking this, a visitor can correct mistakes and immediately update the page. This has the postive effect of a slow reaching of aggreement on the content of a hard to read handwritten letter. It is however occassionally vandalised, this is particalarly the case during the academic year as high school teachers encourage their students to use our material for school projects. We try to check all new edits within 24hrs. All edits are logged and checked by staff. A teacher can keep a eye on recent edits here on a live page and amend when appropriate.

robots & crawlers

Please obey our robots.txt file. We try to treat all robots the same whether they be from a large search engine capable of sending this site significant traffic or a research student getting automated data from our service. We have no hidden material on this website and so we only use this file to keep search engines clean of material underconstruction. This file is also used to direct website crawlers to several sitemaps.

The main current sitemap (sitemap.xml) is generated with google sitemap tool and follows the internet standard containing a 'last modified' time on static html and image files. One sitemap is generated for each lot in the database as a single sitemap would be well above the 50,000 url limit. Currently there are thirty seven lots. A logical link is made available to each of these in our robots.txt file. Also AskJeeves, Bing(MSN), Google and Yahoo! are pinged every 24hrs with updated version. If you work for a search engine and would like us to ping a certain address, please contact us and we'd gladly be of service.

Meta Tags and browser data
Each letter in the database contains hidden format data (including meta tags). Some of this will be useful in catergorising our material. As an example we use the letter 0037-0002 to explain those of interest here. The first line of html code on many letters in the database is

<html LANG="en_uk">

Although most of our letters are written in English, we do try to service as many languages as possible and currently have material in German, French and Spanish. The opening html tag tells which language the letter is written in. This is useful when providing the visitor with links to translation services.

It can also be a great tool for telling the nationality of the author. i.e. An American soldier stationed in London during world war two would have "en_us" while most native Londoners at the same return address would have "en_uk" as above.

<meta name="rating" content="GENERAL">

All letters contain a rating tag. This is used by parent filter software common in schools to prevent youngsters seeing content which may be unsuitable. Also can be used by search engines to tailer their search results depending on the prefer of the searcher. Almost all of the letters and lots have a "GENERAL" rating except those that contain excessive sexual content in which case we use an "ADULT" rating.

Letters involving violence(such as a death threat) is not considered as a seperate rating at this time. This is a continued topic of conversation with teachers using the site as an lesson aid and may change in the near future. We will post it here if it does.

<meta name="date" content="17/05/1858">

The date tag is the date in which the letter was sent i.e. the date on the letter or the date on the envelope of the letter. We will use the UK convention (DD/MM/YYYY) such as below, use four digits for the year and two each for the month and day, placing a preceeding zero when necessary to keep two digits for the month/day.

<meta name="last-edited" content="25/08/2009">

This is the date when the contents of the letter were last modified. This was most likely when a visitor saw a mistake in the transcribed text and corrected it. The date follows UK convention of DD/MM/YYYY, same as above.

<meta name="author" CONTENT="Allen J. Niven">

The author tag is used to display the "author of the letter", that is the person who signed and sent the letter.

<link rel="home" href="http://letter.ie/0037/">
<link rel="prev" href="http://letter.ie/0037/0001.html">
<link rel="next" href="http://letter.ie/0037/0003.html">

For each letter the "link rel" within the head tags of the html document will point to the next/previous letter in cronological order within a particular lot so our visitor can read each letter in a book chapter type way. The "home" will point to the lot page of each letter. For lot pages themselves the next/previous will point to the next/previous lot and home will point to our homepage.

We are interested in adding lots more meta tags in the near future and will be posting how we use them here. In particular we will be adding coordinate data which will allow search engines to display letters that are relavant to the geographical location of the searcher. We are trying to keep things as transparent as possible and if you have any comments or questions, please don't hesitate to ask.