Remove "downsampling" and "upsampling" in doc to avoid confusion

Fix #40

doc/Compare_algorithms.ipynb

@@ -68,8 +68,8 @@
    "outputs": [],
    "source": [
     "ds_in = xe.util.grid_global(20, 15)  # input grid\n",
-    "ds_fine = xe.util.grid_global(4, 4)  # for down-sampling\n",
-    "ds_coarse = xe.util.grid_global(30, 20)  # for up-sampling"
+    "ds_fine = xe.util.grid_global(4, 4)  # high-resolution target grid\n",
+    "ds_coarse = xe.util.grid_global(30, 20)  # low-resolution target grid"
    ]
   },
   {
@@ -166,7 +166,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Down-sampling"
+    "## Increasing resolution"
    ]
   },
   {
@@ -250,7 +250,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "When regridding from low-resolution to high-resolution (down-sampling), `bilinear` and `patch` will produce smooth results, while `conservative` and `nearest_s2d` will preserve the original coarse grid structure (although the data is now defined on a finer grid.). \n",
+    "When regridding from low-resolution to high-resolution, `bilinear` and `patch` will produce smooth results, while `conservative` and `nearest_s2d` will preserve the original coarse grid structure (although the data is now defined on a finer grid.). \n",
     "\n",
     "`nearest_d2s` is quite different from others: One source point can be mapped to **only one destination point**. Because we have far less source points (on a low-resolution grid) than destination points (on a high-resolution grid), most destination points cannot receive any data so they just have zero values. Only the destination points that are closest to source points can receive data."
    ]
@@ -259,7 +259,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Up-sampling"
+    "## Decreasing resolution"
    ]
   },
   {
@@ -320,7 +320,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "When regridding from high-resolution to low-resolution (up-sampling), all methods except `nearest_d2s` produce similar results here. But that's largely because the input data is smooth. For real-world data, it is generally recommended to use `conservative` for up-samplig, because it takes average over small source grid boxes, while `bilinear` and `nearest_s2d` effectively throw away most of source grid boxes.\n",
+    "When regridding from high-resolution to low-resolution, all methods except `nearest_d2s` produce similar results here. But that's largely because the input data is smooth. For real-world data, it is generally recommended to use `conservative` for upscaling, because it takes average over small source grid boxes, while `bilinear` and `nearest_s2d` effectively throw away most of source grid boxes.\n",
     "\n",
     "`nearest_d2s` is again different: **Every** source point **has to be** mapped to a destination point. Because we have far more source points (on a high-resolution grid) than destination points (on a low-resolution grid), a single destination point will receive data from multiple source points, which can accumulate to a large value (notice the colorbar range)."
    ]
@@ -342,7 +342,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.2"
+   "version": "3.6.6"
   },
   "toc": {
    "nav_menu": {},